In section Geographical Distribution the present-day distribution of Gesneriaceae was described, which is well known and no longer poses scientific problems. The subject of this section is to describe the putative phytogeographical history, that is the geographical origin of Gesneriaceae and their splitting and migration events in the course of earth history. As we have to survey a time range of some 50 millions of years, reconstruction is difficult and subject to many uncertainties. Twenty years ago, no sound hypothesis existed as to how the present pattern of Gesneriaceae distribution could be explained. The situation is – based on the molecular methods of phylogeny reconstruction and time dating (see Age of Gesneriaceae and Major Subgroups) – much better today, but still far from satisfactory.

One hypothesis is that the Gesneriaceae family is of “Cathaysian” (mainland Asia, China) origin (H.S. Wang 1989, Ying et al. 1993). This idea was based on the “facts” that China harbours many endemic genera (31 in traditional Chinese taxonomy, e.g. Wang et al. 1990, 1998) and an up to three times higher number of genera as compared to tropical SE Asia. However, this hypothesis does not take into account (a) that the generic concepts of many Chinese genera were very narrow (in fact, the studies of Möller et al. 2011a showed that many small and “endemic genera” must be sunk in synonymy with larger genera such as Oreocharis, Petrocodon, Hemiboea etc.), (b) that the centre of evolutionary diversification was uncritically equated with the centre of origin, and (c) that the hypothesis is based solely on the consideration of Asiatic gesneriaceae, ignoring other parts of the family.

The first to present a more balanced hypothesis was Burtt (1998), who correlated the present-day distribution with geological events such as plate tectonics and continental drift. In brief, the essential points are: (a) Gesneriaceae is an ancient family of Gondwanaland origin. (b) The extant Coronanthereae represent the last survivors of ancestral Gesneriaceae; while some members survived on islands of the Australian plate, another part invaded the Americas via Antarctica and gave rise to the Gesnerioideae. (c) While the South American Coronanthereae became almost extinct (the three monospecific genera Mitraria, Asteranthera and Sarmienta being the last survivors), the Gesnerioideae evolved explosively in the South and Central American tropics. (d) The Australasian Coronanthereae moved northwards on the Indian plate and gave rise to the Didymocarpoideae. (e) On the way north, parts of Didymocarpoideae spread to Madagascar and colonized mainland Africa from there. (f) The Indian plate carried the Didymocarpoids finally to the Asiatic continent. Here an explosive radiation occurred and from here Europe, Sundaland and the Pacific were colonized.

Though Burtt’s hypothesis appears plausible from a geographical point of view, the geological events do not match with the supposed migration events. The breakup of Gondwanaland occurred some 100 mya, while Gesneriaceae originated distinctly later, probably some 50 mya (see Age of Gesneriaceae and Major Subgroups).

Moreover, as is apparent from the molecular data, the closest relatives and descendants of common ancestors of Gesneriaceae (Peltanthera, Calceolariaceae) are all centered in South America. Sanango and Sanangoideae, respectively, is sister to Gesnerioideae plus Didymocarpoideae (= Gesneriaceae in the traditional sense). It is clearly a remnant of the most ancient group in the family and its distribution and there is no evidence that its origin was elsewhere than in South America. Therefore, there is little doubt that Gesneriaceae originated in South America, long after this continent broke and drifted away from Gondwanaland.

The Old World has been reached by three independent groups: (1) the Titanotricheae (Titanotrichum oldhamii), (2) the Coronanthereae (with three genera in South America and 5–6 in the SW Pacific and Australia, (3) the Didymocarpoideae. While we have some rough idea on the times when the origin and successive migration took place, we know almost nothing about the dispersal mechanisms and migration routes.

In particular, it is enigmatic how the Titanotricheae reached East Asia. The hitherto available molecular data suggest that the tribe/genus is part of Gesnerioideae and perhaps closely related to Napeantheae (Perret et al. 2013). So there is some probability that the common ancestor was in South America. Perret et al. (2013) suspect that Titanotrichum arose in East Asia, following long-dispersal from South America during the Miocene.

For the Coronanthereae two migration routes must be taken into consideration: (1) direct (trans-Pacific) migration (by long-distance dispersal) and (2) migration via the Antarctica. Though the first “involves long distances, even for island hopping, and is against current prevailing winds, and Gesneriaceae is absent from the SE Pacific”, Woo et al. (2011: 249) reached on grounds of geological-time considerations the conclusion: “As unlikely as trans-Pacific dispersal may seem, migration via Antarctica seems less likely.” Importantly, these authors also provide evidence that migration of the Coronanthereae to the Old World was not a single event, but occurred in at least two waves.

With respect to the largest group, the Didymocarpoideae, we are completely ignorant how these reached the Old World.

Epithema and Rhynchoglossum (Epithemateae) have a transcontinental distribution, reaching West Africa and the New World, respectively. While Burtt (1998) considered this as an indication of a particularly high geological age, it is more plausible to explain the disjunct distribution of individual species by rather recent expansion or introduction events than by continental drift. The close relationship of Rhynchoglossum azureum with South Indian species (Mayer et al. 2003) suggests that the species has reached America rather recently, perhaps via transpacific trips or migrations of early Polynesians. This is also suggested by the current localities which are usually near former population centres of ancient dwellers along the Pacific coast.

The “Didymocarpoids” (tribe Trichosporeae sensu Weber et al. 2013) are the largest group of subfam. Didymocarpoideae, exhibiting a huge distribution area that includes southern Europe, Africa, Malesia and the Pacific. The molecular-phylogenies of Möller et al. (2009, 2011a) demonstrate clearly that all clades diverging at the base are from mainland Asia (S India to S China). This is probably the place of arrival and first center of phylogenetic radiation. From there, some groups migrated to Africa and to Europe. In a final step, the groups remaining in Asia experienced an explosive radiation, with dramatic expansion to the South East Asian islands, Australasia including New Guinea, and the Pacific.

These general considerations have been recently backed up by the use of modern molecular-biogeographic analysis methods, see Roalson & Roberts (2016), particularly the section “Biogeographic analyses”.