SummaryIt is estimated that around half of all species of flowering plants show self‐incompatibility (SI). However, the great majority of species alleged to haveSIsimply comply with ‘the inability of a fully fertile hermaphrodite plant to produce zygotes when self‐pollinated’ – a definition that is neutral as to cause. Surprisingly few species have been investigated experimentally to determine whether theirSIhas the type of genetic control found in one of the three established mechanisms, that is, homomorphic gametophytic, homomorphic sporophytic or heteromorphicSI. Furthermore, our knowledge of the molecular basis of homomorphicSIderives from a few species in just five families – a small sample that has nevertheless revealed the existence of three different molecular mechanisms. Importantly, a sizeable cohort of species are self‐sterile despite the fact that self‐pollen tubes reach the ovary and in most cases penetrate ovules, a phenomenon called late‐acting self‐incompatibility (LSI). This review draws attention to the confusion between species that show ‘self‐incompatibility’ and those that possess one of the ‘conventionalSImechanisms’ and to argue the case for recognition ofLSIas having a widespread occurrence and as a mechanism that inhibits selfing and promotes outbreeding in many plant species.ContentsSummary717I.Introduction717II.Conventional self‐incompatibility mechanisms718III.The occurrence of ‘conventional self‐incompatibility’ in angiosperm taxa718IV.A molecular perspective720V.Species lacking established conventional self‐incompatibility in the angiosperms722VI.Late‐acting self‐incompatibility722VII.General conclusions728Acknowledgements730References730