Re: Distributions of morphological characters and more

[David Marjanovic <david.marjanovic@gmx.at>]

----- Original Message -----
From: "Michael Mortimer" <mickey_mortimer111@msn.com>
Sent: Wednesday, August 17, 2005 7:51 PM

> > Additionally, I've the following character in my analysis:
> > Premaxillary tooth number: 0- four; 1- two; 2- three; 3- five; 4- six;
> > 5- seven; 6- zero.
> >
> > Should the sixth character state (zero premaxillary teeth) be
> > separated as "premaxillary teeth: 0-present; 1- absent" or is it fine
> > as a state within a character counting the number of premaxillary
> > teeth?
>
> I separate it in my coelurosaur analysis, and code taxa without 
> premaxillary
> teeth as inapplicable for tooth number.

Fortunately TNT can cope with inapplicable characters...

> That's only because I order the tooth number character, and toothlessness 
> seems to be able to evolve from both reducing tooth number (e.g. 
> oviraptorosaurs) and increasing it (e.g. ornithomimosaurs).  So I don't 
> want to favor or punish either possibility.
> But if you left it unordered, I don't see why you'd need the second
> character.

Why not make a stepmatrix. Order "two -- three -- four -- five -- six --  
seven", and give the step from each of those to "zero" (or, alternatively, 
just those from "two" and from "seven" to "zero" -- for example) a cost of 
1. PAUP* has no problem with this, plus it spares you the problem of 
inapplicability with which PAUP* can't cope.

I wouldn't completely unorder the character.