Article: Mathematician solves evolutionary mystery???

[nephilimiyr]

http://www.sciencedaily.com/releases/2003/09/030929055601.htm

Very interesting article that may either solve some of the questions about evolution or add more questions to it.

exerpt:

In other words, why do some descendants of a parent species evolve hundreds of different species, while others produce so few they seem to be practicing family planning?

 

[lucaspa]

http://www.sciencedaily.com/releases/2003/09/030929055601.htm

Very interesting article that may either solve some of the questions about evolution or add more questions to it.

I can see there are going to be some problems with this paper. BTW, I now have the PDF of the entire paper, not just the news headlines.

What Pinelis is trying to do is find a mathematical function that will fit the distribution of evolutionary trees. This is not a cause of that distribution, but a description.

For instance, Penelis ignores environment and selection. He assumes that speciation is a random event. Biologically, that ain't true. It may be applicable only in the sense that, with two species occupying the same area, it may be "random" that one speciates relative to the other.

In the news article, the data of "living fossils" he is using not good. Coelencanths are a genus, not species. Right now there are two species of coelencanth alive, and neither are the species that was alive at the end of the Cretaceous.

From the paper:
"Note that the term ‘random’ is used in this paper in the general sense assumed in probability theory; in particular, this term does not necessarily imply that the corresponding probability distribution is uniform in some sense. In most contexts in this paper, the term ‘random tree’ will actually mean ‘the tree effected by a random evolutionary mechanism’.
Let us assume that the evolution begins with a single species, which is at the root of the random tree 4.
Assume that any species may give birth, if ever, only to one species at a time."

The last assumption is probably not valid.

As I look at the paper, the mathematics is quite fun, but the biology is lacking. For instance, for "speciation" Penelisi is considering only cladogenesis. He ignores anagenesis. We'll have to see what happens.

 

[nephilimiyr]

So what your saying is, he really hasn't proven anything, just yet?

 

[lucaspa]

So what your saying is, he really hasn't proven anything, just yet?

Right.

What he has is an interesting idea: the observation that some lineages have lots of species while other lineages have few species is due simply to probability. Not some deep underlying biological principle of structure of the genome, etc.

That is, take two species A1 and B1 that occupy roughly the same ecological niche in the same geographical area. If you view speciation as a "random" event, then one of the species will split in two and give rise to a new species, call it A2 (so we know it is in the A lineage). Now you have three species in the same geographical area in roughly the same ecological niche. Now once again you have 1 of the 3 species give rise to a new species -- randomly. It could be either A1, A2, or B1. Well, the odds are 2:1 that it will be one of the species in the A lineage, right? So now we have species A3.

A1, A2, A3, and B1 are now present. Again there is a "random" speciation event and the odds are now 3:1 that it will be one of the A lineage.

See where this goes? Lineage A keeps adding new species "randomly" simply because the odds favor that a new species will be from lineage A after that first speciation. Let's say at this stage B beats the odds and you get species B2. The next stage still has odds of 3:2 that the next species will be from lineage A.

Now, my problem is that Penelis ignores biology. Species just don't split "randomly". Either you have geographical separation or ecological separation. Otherwise, same population in the same area in roughly the same ecological niche means you have gene flow to keep the population one species. That population may transform over time to a new species -- anaganesis - but it won't split.

In order to split a species --cladogenesis -- you need disruptive selection by geographical or ecological separation. Now, it may be that for sympatric speciation the movement to a new ecological niche in the same geographical area is "random". That is, it is random which of the two populations has members that move to the new niche. Penelis' work would then imply that most speciation is sympatric. However, the fossil record indicates that most speciation is allopatric -- geological separation.

My guess is that evolutionary biologists are simply going to ignore the paper because the math, altho interesting, simply doesn't reflect the biology. I could be wrong. We'll simply have to wait and see.