An attempt is made to estimate Lagrangian time scales connected with clouds in the atmosphere. Three-dimensional wind and relative humidity obtained from the First GARP Global Experiment (FGGE) data set are used to compute three-dimensional trajectories with accompanying clouds. From the meteorological history of the trajectories, estimates are made of the average time from the release of a trajectory until the first cloud passage and the average time between first and second cloud passage. Geographical and seasonal variations of these parameters are calculated and discussed. Since the ''clouds'' parameterized from the FGGE data are limited to those connected with the large-scale flow in the atmosphere, a substantial bias in our estimates is introduced. This implies that the estimates are more representative for major cloud systems, often connected with precipitation, than for clouds in general.