Abstract. Mean properties of individual contrails are characterized for a wide range of jet aircraft as a function of age during their life cycle from seconds to 11.5 h (7.4–18.7 km altitude, −88 to −31 °C ambient temperature), based on a compilation of about 230 previous in situ and remote sensing measurements. The airborne, satellite, and ground-based observations encompass exhaust contrails from jet aircraft from 1972 onwards, as well as a few older data for propeller aircraft. The contrails are characterized by mean ice particle sizes and concentrations, extinction, ice water content, optical depth, geometrical depth, and contrail width. Integral contrail properties include the cross-section area and total number of ice particles, total ice water content, and total extinction (area integral of extinction) per contrail length. When known, the contrail-causing aircraft and ambient conditions are characterized. The individual datasets are briefly described, including a few new analyses performed for this study, and compiled together to form a contrail library (COLI). The data are compared with results of the Contrail Cirrus Prediction (CoCiP) model. The observations confirm that the number of ice particles in contrails is controlled by the engine exhaust and the formation process in the jet phase, with some particle losses in the wake vortex phase, followed later by weak decreases with time. Contrail cross sections grow more quickly than expected from exhaust dilution. The cross-section-integrated extinction follows an algebraic approximation. The ratio of volume to effective mean radius decreases with time. The ice water content increases with increasing temperature, similar to non-contrail cirrus, while the equivalent relative humidity over ice saturation of the contrail ice mass increases at lower temperatures in the data. Several contrails were observed in warm air above the Schmidt–Appleman threshold temperature. The emission index of ice particles, i.e., the number of ice particles formed in the young contrail per burnt fuel mass, is estimated from the measured concentrations for estimated dilution; maximum values exceed 1015 kg−1. The dependence of the data on the observation methods is discussed. We find no obvious indication for significant contributions from spurious particles resulting from shattering of ice crystals on the microphysical probes.