An inventory of the species in and around the clove agroforestry systems (AFS) on the east coast of Madagascar revealed that none of the currently grown species derive from the native Malagasy biodiversity: the herbaceous and woody plants which comprise planned agrobiodiversity, i.e. are deliberately planted by the farmer, are without exception, introduced species. The associated agrobiodiversity, including plants that grow spontaneously and are protected or tolerated by farmers, is composed of a mix of native and introduced species. This situation, which is paradoxical for an island with a very high level of endemicity is the result of a double process of species introduction. The first is associated with the waves of migration, which, during the past three thousand years or so, allowed colonisation of the island, previously devoid of human occupants. The migrant populations originating from Austronesia and then from Africa, introduced their own crops and associated growing practices and only exploited local biodiversity through gathering, which did not require planting or tending. This situation is therefore the result of a long process of farmers' strategies to manage natural resources in a context of subsistence agriculture. The second process is more recent and is associated with increasing occupation by Europeans from the 16th/eighteenth century onwards, which peaked at the end of the nineteenth century with the French colonisers, who introduced and then developed cash crops (coffee, cloves, vanilla). This long-term retrospective study sheds light on the dynamics at the origin of the biodiversity "melting pot" visible today in the AFS along the east coast of Madagascar. It shows the constant evolution and adaptation of farmers' strategies and cropping systems, but also their historical roots. The cohabitation between introduced and indigenous species, observed today in Madagascar, is found in other AFSs built around an introduced cash crop, in various tropical zones. Knowing the origins and determinants of this cohabitation could be useful to support future evolutions of these cropping systems in the current context of changing agricultural practices in response to ecological and macroeconomic pressures.