> Are you ignoring what dave said

> Morphometry can be used ....... to classify unknown samples into racial 
> groupings,

I hope not.  Dave wrote a carefully crafted, carefully qualified appraisal 
of the abilities and limitations of the technique, in his (IMO expert) 
opinion.

Unfortunately, in quoting his words, you left out the qualification, and 
then, later, also omitted the subsequent discussion by him of those 
qualifications.  To me these were the most meaningful parts, and exactly the 
reasons I am skeptical about many applications (misapplications) of 
morphometry.

After all we all know that morphometrics has good uses--Dave pointed them 
out--but many tend to ignore the limitations of this tool, and try to use 
morphometrics to reach conclusions that require greater precision than may 
be possible, due to the difficulty in calibration--in absence of known and 
double-verified samples of the whole population (or populations) being 
examined, or even a full understanding of the entire make-up of those 
populations

Dave said, "within a *closed* population (emphasis mine) or to classify 
unknown samples", also, "Any morphometry project is only as good as the 
database that it is founded upon", and "if you are trying to identify 
something like AHB, you will need a database that has been developed from 
colonies that are known to be AHB by behaviour that can be traced back to 
Kerr's experiment".

> Isn't what ARS were doing just what Dave has said was possible to be done? 
> Screening an unknown sample into a racial group?

Maybe, but I have strong reservations about the reliability of such an 
effort.

What I wonder about, and what made me a bit surprised, was that I really 
doubt that Dave's conditions have been met.  Does such a database exist, and 
how good is it?

I leave it to you and others to re-read what Dave said, and perhaps Dave 
will say if he thinks I misundertood or ignored his expanation.  I thought 
what he said was clear enough, and that I understood, but perhaps not.

Thanks for questioning.

allen 

-- Visit www.honeybeeworld.com/bee-l for rules, FAQ and  other info ---