Terrestrial impact rates are calculated for long and short-period comets crossing the earth’s orbit. For a random long-period comet the impact probability is 2.2 × 10−9 per perihelion passage and the most probable impact velocity is 56.6 km s−1. Based on the sample of 20 observed earth-crossing short-period comets the mean impact probability is 6.6 × 10−9 per perihelion passage or 8.2 × 10−10 yr−1, and the most probable impact velocity is 28.9 km s−1. For the long-period comets, based on an estimated flux of 16 comets AU−1 yr−1 brighter than H10 = 11, this leads to an impact rate of 2.9 × 10−7 yr−1 or a cratering rate of 0.06 × 10−14 km−2 yr−1 for craters with diameter ⩾10 km. Assuming a steady state number of approximately 800 short-period comets in earth-crossing orbits leads to an impact rate of 6.6 × 10−7 or a cratering rate of 0.13 × 10−14 km−2yr−1 for 10 km or greater craters. The total cratering rate from both long- and short-period comets is about 15% of the observed rate derived from known terrestrial astroblemes, though there is substantial uncertainty in both estimates. The estimated cratering rate from earth-crossing asteroids (Shoemaker and others, 1979) is about twice the observed rate. Monte Carlo simulations of the dynamical evolution of the Oort cloud give estimates of the cometary flux rate in the early solar system on the order of 200 times the current flux.