Into Africa, from Australia
It took me a long time to get used to the idea that the plants (and algae and animals..) we see today, and where they live, are due to common responses to big events that happened millions of years ago. That is, continents shifting apart and mountain ranges pushing upwards, splitting lots of species in two (or three, or four...) and the isolated subgroups evolving become new species.
It took time because we often (used to) talk about centres of origin or dispersal, as though every species started somewhere and then radiated out into new territory. Although it was accepted species had to, and did, move into new areas from time to time, the important changes were thought to be due to the breaking apart of an established range.
The job for a systematist - someone who studies the relationships between living things - was to look for these patterns in as many organisms as possible. Sure there were likely to be many layers - with historical climate change and a variety of small and large scale 'earth movements' making the signal rather fuzzy at times - but the goal was to find the events that split the range.
More recently there has been a return to 'long-distance dispersal' as a legitimate scientific hypothesis. In the past long distance dispersal was usually evoked without much evidence. In fact, it is usually necessary to discount major geographical (or other) barriers and pattens first before we can consider dispersal.
And this is what the latest paper on the protea family (Proteaceae) has done. It's a very neat piece of work led by Hervé Sauquet from the Jodrell Lab at Kew Gardens, ably assisted by our very own Peter Weston, and Nigel Barker (South Africa), Cajsa Anderson (Sweden), David Cantrill (Melbourne) and Vicnent Savolainen (London).
Sauquet and colleagues wanted to find out more about how the 6,000 species of plants that grow in the Cape Region of South Africa and only there, got there. They chose a subfamily of Proteaceae, called the Proteoideae, to work on - all but one of the 331 species of Proteaceae native to the Cape belong to this subfamily.
They sequenced seven different bits of DNA in representatives of the 13 genera in this subfamily to track the evolutionary relationships. Importantly they also measured and analysed the pollen (coming up with 22 characters) of present-day plants, as well as well-dated and reliable fossils of plants that form part of this subfamily's 'family tree'.
What they found is that one major chunk of the subfamily - including the well known genus Protea - has its origins back about 70 million years ago, when it appears to have split from a close relative in New Caldedonia. This date is too young to be explained by Africa pulling away from New Caledonia, so it is hypothesised that there was at least one 'long-distance dispersal event' (i.e. a seed or a bit of a plant travelled a long way!) into Africa.
The other major group has even more recent origins. This one has links to the Australian genus Adenanthos, and is thought to have begun its separate evolution about half as long ago (39-46 million years in the past). This is way too young to have started with the separation of Africa and Australia as Gondwana broke apart (the last connection was severed around 105 million years ago), and again a long-distance dispersal event is invoked. In this case the authors note the small seeds would be readily dispersed by wind so it's not unreasonable.
A further important divergence in the subfamily, this time between the Australian Petrophile and some South African relatives, is somewhere in between the other two 'events', and a third you-know-what is proposed, again supported by the seeds being capable of wind dispersal.
There is more detail in the paper and I'm probably over generalising in my summary, but the basic message is that there is more than one point of origin for this group of South African plants; long-distance dispersal is the most likely starting point in each case; and the wonderful diversity of Proteoideae on the Cape has its origins in Australia and New Caledonia.
If you are worried about, or interested in, the assumptions on which these conclusions are based, do take a look at the paper in Molecular Phylogenetics and Evolution 51: 31-43 (2009) - although it does cost to download. The authors were careful to exclude comparisons with analyses that dated their evolutionary trees with either a single fossil or only using a strict molecular clock.