outgroup versus ordered characters in Cladistics

[Jong, R. de]

In my opinion there is no connection between outgroup and ordering of
characters. In parsimony analyses the numbers of changes from one character
state to another are counted. In unordered characters the change from any
one state to any other state is counted as one step.  Thus, in an unordered
character with the states 0, 1 and 2, the change from 0 to 1 is one step,
from 0 to 2 is one step, from 2 to 1 is one step, etc. In an ordered
character the number of steps is the absolute value of their arithmetic
difference. A change from 0 to 2 is, thus, counted as two steps, from 1 to 5
as four steps, from 4 to 1 as three steps etc. From this it can easily be
seen that the order is irrelevant in a two-state character, the difference
always being 1. Hennig86 and PAUP do not differ with regard to the way
ordered and unordered characters are treated, in both one can interactively
change characters from ordered to unordered and vice versa.
An outgroup is only added to the matrix to root the tree. It determines the
direction of the character changes. The number of steps in the ingroup
remains unchanged. If the states of an (ordered or unordered) character are
coded 0, 1, 2, 3, etc., it does not follow that the 0 state is
plesiomorphic. Only if the same state of the same character in the outgroup
is also coded 0, there is a chance that the 0 state is plesiomorphic.
However, the terms plesiomorphic and apomorphic pertain to a sistergroup
relationship. They indicate which character state was taken over unaltered
from the ancestor, and which was altered. What is apomorphic at a particular
level in the tree may be plesiomorphic at a higher level (i.e. towards the
crown). If a character state is 0 in the outgroup (and thus plesiomorphic at
the level of ingroup-outgroup), and it occurs higher up in the tree as a
result of a reversal, it is apomorphic at that particular level.

With kind regards,

Rienk de Jong

> -----Oorspronkelijk bericht-----
> Van:  Olivier De Clerck [SMTP:Olivier.Declerck at RUG.AC.BE]
> Verzonden:    dinsdag 22 juni 1999 12:22
> Aan:  Multiple recipients of list TAXACOM
> Onderwerp:    outgroup versus ordered characters in Cladistics
>
> Dear all,
>
> Recently we were wondering in the department about the difference
> between defining an outgroup in a parsimony analysis and explicitely
> stating that the characters are ordered.
>
> If one creates an outgroup in  a parsimony analysis, which has a
> 0-value for a two-state character, does this automatically mean that
> the 1-value is the apomorphic state and 0-value the plesiomorphic
> state? I would presume it is. BUT what is in this case the additional
> value of defining that the characters are ordered (0-value =
> plesiomorphic ; 1-value = apomorphic)?
>
> Additional question: is there a difference between Hennig86 and Paup
> on this point?
>
> Looking forward hearing from you,
>
> Olivier De Clerck.
>
>
> =================================================
> Dr. Olivier De Clerck
> Dept. Biology
> K.L.Ledeganckstraat 35 - B-9000 Gent, Belgium
> tel 9-264.50.64 / fax 9-264.53.34 / olivier.declerck at rug.ac.be
>
> ==================================================