Re: Erketu ellisoni online free access

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

> The data matrix has more question marks than known character states.

Not true. These are just the codings for *Erketu*. The rest is taken 
unchanged from Wilson (2002), except for one character state of one taxon 
explained in the text.

> Nevertheless, the results pretend that we know something about the
> question marks (as if they weren't there).

How so? At worst it lets us apply parsimony to infer what the true states 
are -- _after_ the tree has been made.

> There is nothing wrong in science to
> say "I don't know" than to pretend that we do.

But we do know something. We know that, according to that analysis, 
*Erketu*(*) is with quite some certainty a somphospondyl. The bootstrap 
value of that clade is 89. We can also be fairly sure, though less so, that 
*Erketu* is not a member of the smallest clade that contains *Malawisaurus* 
and *Alamosaurus*, because that clade has 79 % bootstrap support. These 
values are better than almost everything inside the latter clade (51, 56, 
60, 70, 81 and 83) and much better than the, one should think, quite obvious 
Neosauropoda (less than 50 % -- not shown).

But then I have my doubts about Somphospondyli. It seems to be held together 
mainly by its somphospondylous vertebrae... right?

* Meant to be pronounced the French way. The etymology section says "Erketü" 
and "Yesügei". -- BTW, in spite of this the paper is totally full of typos. 
Some are consistent throughout the text and references, so may be genuine 
errors, such as "Tshuihiji" and "*Mamenchisaurus houchuanensis*".