Cape diversification and repeated out-of-southern-Africa dispersal in paper daisies (Asteraceae–Gnaphalieae)

Abstract

The large daisy tribe Gnaphalieae occurs in extra-tropical habitats worldwide, but is most diverse in southern Africa and in Australia. We explore the age and evolutionary history of the tribe by means of a phylogenetic hypothesis based on Bayesian analysis of plastid and nuclear DNA sequences, maximum likelihood reconstruction of ancestral areas, and relaxed Bayesian dating. Early diversification occurred in southern Africa in the Eocene–Oligocene, resulting in a grade of mostly Cape-centred lineages which subsequently began speciating in the Miocene, consistent with diversification times for many Cape groups. Gnaphalieae from other geographic regions are embedded within a southern African paraphylum, indicating multiple dispersals out of southern Africa since the Oligocene to Miocene which established the tribe in the rest of the world. Colonisation of Australia via direct long-distance trans-oceanic dispersal in the Miocene resulted in the radiation which produced the Australasian gnaphalioid flora. The similarly diverse regional gnaphalioid floras of Australasia and southern Africa thus exhibit very different temporal species accumulation histories. An examination of the timing and direction of trans-Indian Ocean dispersal events in other angiosperms suggests a role for the West Wind Drift in long-distance dispersal eastwards from southern Africa.

Introduction

The daisy tribe Gnaphalieae is represented on most continents but forms a significant component of dry- and cool-temperate floras in the Southern Hemisphere. This tribe of ca. 1240 species (Bayer et al., 2007) has two main centres of diversity: southern Africa, where one quarter of the ca. 2200 Asteraceae species are gnaphalioids (Koekemoer, 1996) and Australasia with nearly 500 species (Bayer et al., 2007). The remainder are found in South and Central America (ca. 100 species; Anderberg, 1991), the rest of Africa, the Mediterranean, Asia, and North America (Bayer et al., 2007).

In the morphological cladistic analysis of Anderberg (1991), the early-diverging lineages consist of taxa from both southern Africa and Australia. Bayer et al.’s (2000) molecular phylogenetic study did not include any Australasian taxa but indicated that the earliest diverging lineages contain southern African species from the near-endemic Cape subtribe Relhaniinae. This paraphyletic (Bayer et al., 2000) subtribe was identified by Linder (2003) as the tenth-largest Cape Floral Clade. Elucidation of the phylogenetic relationships in Relhaniinae is clearly key to understanding the evolutionary history of the tribe as well as to providing insight into the origins of the unique Cape Flora of southern Africa. Here we build on the analyses of Bayer et al. (2000, 2002) by constructing a phylogenetic hypothesis for the Gnaphalieae with special focus on members of the Relhaniinae, but including non-relhaniinoid taxa from southern Africa, Australia and the Northern Hemisphere. We ask, firstly, what such a phylogeny can tell us about the biogeographic history of the tribe and specifically, given that the two major centres of diversity are on either side of the Indian Ocean, about the timing and direction of trans-oceanic dispersal(s) in Gnaphalieae. Secondly, we ask whether there is a general pattern in trans-Indian Ocean plant dispersals, and what mechanisms could underly such a pattern.