fortuner at WANADOO.FR
Mon Mar 24 06:51:28 CST 1997
Robin Panza wrote:
>I can see some value in this, but not a tremendous amount. I assume Renaud
>would like to check the data to see if they support the conclusions. How many
Actually, this is only one of the things you could do with the data.
You could also use it with particular identification methods, to do
genus revisions, and every time you need the descriptions of
individual specimens for existing (i.e., published) species.
At the present, the only way to do this is to get the data the hard
way (getting specimens, measuring, etc.). How many of us truly have
the time to redo all this, on top of all the reading and research (and
other responsibilities such as teaching or curating) that we are
What Robin is saying about the possibly poor quality of published data
is true, but I have seen several articles in Syst. Biol. where one
author used the matrix published by another author to show that the
original analysis and conclusions were all wrong. I am only
suggesting to go up one step and start from the data that was used to
create the matrix.
Robin Panza raised other points that would all deserve a long answer.
Here are one-liners:
- The accuracy of any piece of data can be assessed automatically and
objectively by what I call an endorsement function.
- How large is a large projection can be estimated by a fuzzy
- I was not thinking of data not used in the study because they are
aberrant or messed up, I was talking of the good data (specimen
descriptions) that were used in the original study but not published.
More information about the Taxacom