The validity of cladograms (was Re: giant birds)

[Kendall Clements <k.clements@auckland.ac.nz>]

Samuel Barnett wrote:
 
> I am interested by this notion of cladograms being untestable as it looks to
> me like we have a control test with extant organisms.  Ie, get a bunch of
> skeletal experts to draw up a cladogram as they would with fossils, but
> based on living animals.  Then get a dna comparison of all the species
> involved and compare the more reliable dna test with those of the cladogram
> to see if they compare.  Has this or anything like it been done?  BTW, I'm
> assuming the dna test is more reliable than cladistics, otherwise:
> 
> Barrister:    Can you prove that this man is the father of your child?
> Plaintiff:      Yes sir, I got a palaeobiologist to look at their bones and
> it's conclusive...

The answer is an overwhelming yes!!! In the majority of cases 
cladograms of extant organisms based on morphological characters are 
congruent with those based on morphologicl characters. I have mentioned 
this a couple of times to support my view that cladistics is the best 
method we have for reconstructing the phylogenies of extinct organisms. 
Your message implies that phylogenetic analysis of DNA is somehow 
"different" to that of morphological characters. Molecular characters 
(DNA) can be analysed using exactly he same methods as morphologicl 
characters (i.e. parsimony). Molecular characters can also be analysed 
using a modelling approach (i.e. maximum likelihood), and by a 
clustering approach (neighbour joining). The latter is not usually 
considered reliable enough on its own for evolutionary reconstructions.

So what you are really talking about is two issues: (i) Is cladistic 
methodology reliable, and (ii) Are trees based on cladistic analyses of 
morphological characters reliable? I would say the answer to both is 
yes, in MOST cases. To go into detail on (i) the respective merits of 
methods for sequence analysis, and (ii) the merits of morphological vs 
molecular characters would surely try the patience of most list users!

Kendall

----------------------
Kendall Clements
k.clements@auckland.ac.nz