Frontiers of Biogeography Early recognition by Ball and Hooker in 1878 of plant back-colonization (boomerang) events from Macaronesia to Africa

Author(s): Fernandez-Palacios, Jose Maria; Whittaker, Robert J. | Abstract: Recent work in island biogeography has shown that back-colonization (‘boomerang’ events) from islands to continents have occurred more frequently than previously understoodWe report possibly the earliest inference of this pattern, by John Ball and Joseph Dalton Hooker in a book published in 1878.


Early recognition by Ball and Hooker in 1878 of plant back-colonization (boomerang) events from Macaronesia to Africa
Within island biogeography, it has traditionally been believed that whereas mainland to island colonization is key to stocking oceanic islands with species, reverse colonization events from islands to continental mainlands are so uncommon as to be without wider significance (Whittaker and Fernández-Palacios 2007). However, the completion in recent decades of sound phylogenies incorporating almost entire sets of insular and continental related taxa has revealed a surprising frequency of instances of continental populations or species that are intermingled within insular clades, or which are derived from insular ancestors ( Figure 1).
Accepting these phylogenies as working hypotheses of the true evolutionary relationships allows the inference of reverse colonization events, which are also sometimes referred to as retro-colonization, back-colonization, or boomerang events (Caujapé-Castells 2004, 2011, Heaney 2007, Bellemain & Ricklefs 2008, Laenen et al, 2011. While it has also been hypothesized that islands have provided stepping stones for cross-ocean colonization of continental mainlands (as reviewed e.g., in Whittaker et al. 2017), the terms 'back-colonization' and 'boomerang events' specifically imply that species that have diversified in islands from continental ancestors disperse back to the original continent (e.g., Mort et al. 2002, Carine et al. 2004, Laenen et al. 2011.
While the available evidence tends to support progression-rule patterns of colonization from older to younger land-masses as being far more common, evidence of back colonization within archipelagos and between archipelagos and continents has gradually accumulated. Today we know about many boomerangs from Macaronesia to Africa and Europe, but also elsewhere, that have happened across multiple taxa (Mort et al. 2002, Bellemain & Ricklefs 2008, Caujapé-Castells 2011.
But when did the idea of island-to-continent retro-colonisations first emerge? By chance, inquiring in old texts we have found that the idea was already in use in the late 19 th century for explaining the same unexpected pattern of the presence of a few species with Macaronesian characteristics in Moroccan territory. "Amongst the exceptional cases to continental proximity being accompanied by close botanical relationship is the Flora of the Canarian Archipelago, which differs so greatly from that of the northern part of its neighbouring continent, namely from that of Marocco, that it demands notice in any work treating of the vegetation of the latter country.
This diversity between the Maroccan and the Canarian Flora has been pointed out in John Ball's (1878) ‛Introductory observations to the Spicilegium Florae Maroccanae,' where it appears that whilst Marocco, out of 1,627 species of flowering plants, contains 165 endemic plants, it has only 15 that are confined to it and to the Canaries or to it and Madeira. And Ball goes on to remark (p. 301), in respect of these few species in common to both Floras: 'I think it is safe to say that the facts rather tend to show the accidental diffusion of a few Macaronesian species on the adjacent coast of Africa, than to indicate the direct connection between the continent and those islands within a geological period at all recent.' " [our emphasis in bold].
 Insert Fig. 1 around here Consulting Ball's work, cited in the above extract, it is evident that of the 15 species, all but one of them are Canarian species that are also found in the coastal region of South Morocco, the other being common to Madeira and West Morocco (Table 1). Hooker went on to set out possible explanations for the common elements, including (i) extensionism (vicariance by the loss of a land-bridge connection), which he largely dismissed due to the depth of the ocean floor between the Canaries and west Africa, and (ii) the role of fishermen from the islands visiting points on the opposite coast. Of greater interest in the present context is the following extract, which in his cautious way, he presented as one of the possibilities: "We may believe in the trans-oceanic migration of some African species to the nearer islands, along with the transport of some Canarian species (those enumerated in p. 416, and others which may be hereafter found) to the neighbouring continent." (p. 418).
This is to our knowledge the very first explicit account of an island-to-continent retro-colonization or back-colonization event, although this is not to rule out the possibility of an earlier mention that has yet to be rediscovered.
Demonstration of persuasive evidence for Macaronesian to North African retrocolonisation had to wait for more than a century, until the development of molecular tools enabled the construction of sound phylogenies that could confirm it. It was 124 years later when Mort et al. (2002) first described, for the Macaronesian genus Aeonium (Crassulaceae), derived from a Sedum-like ancestor, which colonized the Canaries ca. 15 Ma (Kim et al. 2008), the existence of two boomerangs in North Africa (A. korneliuslemsii in Morocco and A. leucoblepharum in Yemen, Ethiopia, Somalia, Kenia and Uganda; Figure 1). These findings in fact led to the loss of the status of Aeonium as an endemic Macaronesian genus!
José María Fernández-Palacios 1 & Robert J. Whittaker 2, 3 Figure 1. Boomerang events. a) Schematic representation of a retro-colonization or boomerang event. A continental ancestor (A) colonizes the island (colonization, depicted with a red arrow), where with sufficient time and isolation, it radiates, giving rise to four insular neoendemic species: B, C, D and E. One of them (D) jumps back to the continent from which the clade's ancestor originated (boomerang event, depicted with a blue arrow). b) Aeonium phylogeny based on a cpDNA/ITS data set, which includes two African boomerangs within a Macaronesian clade: Aeonium leucoblepharum (above) and A. korneliuslemsii (below), identified with blue arrows. Part (b) was excerpted from Mort et al. (2002)