AbstractThe rainfall persistence at Uccle (Belgium) is examined statistically using 75 years of daily observations. Different kinds of distributions expressing this persistence have been fitted to the sequences of dry and wet days: the geometric series, the logarithmic series, the Eggenberger‐Polya distribution, the simple Markov Chain and the Markov Chain of order higher than one.Wet and dry spells are best described by a Markov Chain model of order four. Nevertheless for wet days, the Eggenberger‐Polya distribution provides a good fit to the observed data. For Uccle, the rainfall persistence seems to remain constant after four consecutive days of precipitat