Farewell to species - reticulation

[Pierre Deleporte]

>Andrew Mitchell wrote:
>
>For a cladogram to accurately track a reticulation event it will have to
>show two lineages fusing into one[.]...  Ever seen Hennig86 or PAUP produce
>a tree like [that]?
>
>
>    A   B     C   Z       D
>    \   \    /    |      /
>     \   \  /   __|_    /
>      \   \/   /    \  /
>       \   \  /      \/
>        \   \/       /
>         \   \      /
>          \___\    /
>               \  /
>                \/
>                /
>
>
>Tom DiBenedetto wrote:
>
>No, but Nelson wrote a paper back in the early eighties about how to
>identify possible reticulation events on a cladogram which has the character
>distributions mapped on. The goal was to develop a protocol for placing
>reticulation lines on the cladogram. I dont know whether anyone ever ran
>with that, but it sure made me feel good to know that someone was thinking
>about it, while I went off to find myself some "well-behaved" group.....:)
>

I did see such reticulate cladograms, they are produced by the CTA program
of John Alroy (1995, Syst. Biol. 44, 152-178), which precisely implements
the idea of Nelson 1983 (check for single hybridization events resolving
many "scattered similarities" into homologies... correct, Curtis?).

Anybody cares around there? Do you know this program, did you try it, like
it? It's ""pure cladistics"" anyway (optimizing contiguity of character
states), it just checks for "likely hybridization events" under Nelson's
criterion, and also allows fixes unit taxa to the nodes of the
cladogram-reticulogram (when no autapomorphies justify a terminal
positioning: so much for "cladistics impose dichotomy").

By the way, in the same volume of "Advances in cladistics" 1983, Chris
Humphries wrote a paper about primary data in hybrid analysis, thus the
very first issues of the future journal "Cladistics" dealt with this
problem (so much for "cladists don't care with reticulation"). But the fact
is that this field of investigation still receives little attention, except
perhaps by botanists who can identify hybrids in the caryotypes. Hence the
fate of Alroy's program apparently largely ignored if I judge by this
thread? Nobody needs it, or is there something wrong with this program, or
did the attempts failed in identifying reticulations this way, justifying
Curtis Clark's statement that hybrids generally keep close to one of their
parents?

There is obviously a field for experimental investigations on the line of
"closing the gap between population biology and phylogenetic systematics":
make hybrids and check for the resulting cladogram of the known lineage.
The artificial selection of pidgeons by Darwin and the discovery of the
genetic code was perhaps sufficient to support the idea of "descent with
modification" as the general process in evolution, and this remains the
general assumption supporting phylogeny reconstruction. But there is no
reason to stop there, provided that it can be proved that more complexe
processes (reticulation) leave clear traces behind.

greetings
Pierre





Pierre Deleporte
CNRS UMR 6552 - Station Biologique de Paimpont
F-35380 Paimpont   FRANCE
Téléphone : 02 99 61 81 66
Télécopie : 02 99 61 81 88