The question has often been asked, "How can we be so sure that Cannas originated in South America?" We do know that the first Cannas introduced to Europe were C. indica L., and although they all came to Europe from the East Indies, they originated from the American continent. Charles de l'Ecluse, who first described and sketched C. indica in his Histoire des plantes rare observées en Espagne (history of rare plants observed in Spain), published in 1576, indicates this origin, and states that it was given the name of indica, not because the plant is from India, in Asia, but because this species was originally transported from America: "Quia ex America primum delata sit"; and at that time, one described the tropical areas of that part of the globe as the Western Indies; English speakers still call them the West Indies.
Much later, in 1658, Pison made reference, in his Histoire naturelle du Brésil (natural history of Brasil), to another species which he documented under the vulgar name of 'Albara' and 'Pacivira', and which resided, he said, in the shaded and damp places, between the tropics; this species is Canna angustifolia L., (later reclassified as C. glauca L. by taxonomists).
Without exception, all Canna species that have been introduced into Europe can be traced back to the American continent, and it can be asserted with confidence that Canna is solely an American genus. If Asia and Africa provided some of the early introductions, they were only varieties resulting from C. indica and C. glauca cultivars that have grown for a long time in India and Africa, but not from species growing in a spontaneous state.
The final argument is that Canna seeds have never been discovered by archeologists in the Old World, and the hard shells of Canna would have ensured that some would have survived.