Abstract The comonomer sequence distribution of a propylene–ethylene random copolymer has been investigated by a combination of temperature rising elution fractionation, 13 C nuclear magnetic resonance spectroscopy, Fourier transform infrared spectroscopy and differential scanning calorimetry (d.s.c.). The propylene–ethylene copolymer exhibited a wide range of comonomer compositional heterogeneity and there were no detectable long ethylene sequences containing three or more adjacent ethylene units in any of the fractions. The copolymer was mainly composed of long propylene sequences with an occasional ethylene unit. Sequences such as PPE, EPE, and PEP were present. The sequence distributions of all the fractions did not fit either Bernoullian or first order Markovian statistics. The ethylene content and the comonomer distribution had a marked effect on the crystallisation kinetics and melting behaviour of the fractions.