[Taxacom] Dispersal... again

[Michael Heads]

Hi Karl,
?
The standard model suggests that a group?can't survive as a metapopulation crossing 50 km between islands, now or in the past, but?was able?to cross thousands of kilometers from the mainland.?There seems to be a problem.
?
The idea (Clague, 1996) that there was a period of no islands (or no high islands) in the Hawaii chain is?not irrelevant, it is a? critical part of the standard model. This?is why everyone cites it. If there wasn't a period of no islands,?groups could have survived on the Hawaiian chain indefinitely, hopping from older islands to new ones.?Clague and others have emphasized the potential sources of error in their calculations, but these have generally been ignored.
?
You refer to?'the most wildly optimistic past scenario of islands scattered?across the Pacific'. ?Don't you accept that all the thousands of atolls and?flat-topped seamounts such as the Musicians were once high islands? Why have Hawaiian biogeographers?never cited groups such as the Musicians???
?
The process termed 'Long distance dispersal' does not require long distance, which is why chance dispersal or jump dispersal may be a more accurate term. The process is often used to explain differentiation across a small-scale barrier, e.g. a river. For example, even if the New World was only 20 km (or 100 meters) from the Old World, a model of chance dispersal to explain a break between New World and Old World sister groups would be unnecessary. 
?
You suggest it is 500 miles from Hawaii to the?nearest seamount that could have been a source. But that only considers?the current seamounts. For example,?there could have been?prior islands now buried under younger?volcanics. Even in the Hawaiian islands?there is plenty of evidence for renewed volcanism?(not predicted in pure?hotspot theory). But even if there weren't?and we rely on existing seamounts 500 miles away, the distance isn't?relevant. Simple, everyday movement using normal means and without differentiation is not chance dispersal.?For example, a?metapopulation survives?by normal movement, and does not involve?chance dispersal.?The standard model for Hawaii proposes?that the groups there dispersed from the mainland by chance dispersal, and then differentiated on Hawaii.?The Hawaii clade is different from the mainland one because of chance.??But what are now Hawaiian endemics?could have?evolved on, say, the Musicians
 seamounts or the Line Islands, dispersed to Hawaii (by normal movement, using normal means and without differentiation), and then gone?extinct on?the Musicians and Lines?as they sank. This is a?very different model and does not involve chance dispersal. There is excellent evidence of high-island, wet forest groups formerly present on what are now atolls (e.g. mollusc fossils) so why are these islands not?considered as?potential?sources for Hawaiian biota??Why do groups always have to come from the mainland or high islands thousands of kilometers away, just because their current sister groups are there?
?
The Line Islands form an archipelago of atolls that stretches 4000 km between?Hawaii and French Polynesia. The atolls?are former?high islands erupted in the?Cretaceous.?There are now numerous molecular studies showing sister groups in Hawaii and French Polynesia, but the pattern is?always attributed to long distance, chance dispersal between the two areas?and the Line Islands are never mentioned. Very strange.

Michael

Wellington, New Zealand.


My papers on biogeography are at: http://tiny.cc/RiUE0


________________________________
From: Karl Magnacca <kmagnacca at wesleyan.edu>
To: Michael Heads <michael.heads at yahoo.com>
Cc: "taxacom at mailman.nhm.ku.edu" <taxacom at mailman.nhm.ku.edu>
Sent: Friday, 28 October 2011 1:41 AM
Subject: Re: [Taxacom] Dispersal... again


On Wed, October 26, 2011 2:11 pm, Michael Heads wrote:
> Hi Karl,
>
> You don't need continuous land (e.g. a lost continent) for
> vicariance. For example, two metapopulations on an active island
> arc (surviving in situ on individually ephemeral islands) can be
> separated by rifting.

They can be, but it depends on the organism.? For many things that
have speciated on islands, even crossing a 50-mile channel between
islands is such a rare event that they don't exist as
metapopulations between oceanic islands.

> The dispersal model for Hawaii relies entirely on the suggestion
> that there was a period (33-29 Ma) in which no islands (Clague et
> al, 2009:1) or no high islands (Clague et al., 2009: 10) were
> emergent, but the calculation this was based on is dubious.

No, this is an interesting part of the story but irrelevant.? Even
granting the most wildly optimistic past scenario of islands
scattered across the Pacific, you're still left with having things
cross 500 miles of ocean from the nearest seamounts that could
possibly have ever been high islands with any significant diversity.
By my definition, that still counts as long-distance dispersal.

Also, the 2009 paper was based on coral growth, not the approximate
calculated height.? On top of that, even that number is the maximum
height attained, which doesn't take into account what is removed by
erosion.? Oahu and Kauai have a long way to go before submergence,
but they've already had most of their tops removed so that only the
core of the volcanic pipes remain at the highest points.? So if
anything, Clague's 1996 underestimate is probably an *overestimate*
of the relative height between the original shoreline and the peak
at submergence.

Karl