Hallo, dear morphometters, I need an advise of yours on the subject
which lies a little bit aside of the topics of this forum.

I study structure of within-species variation based on the euclidian
metrics. The distance matrices are routinally represented graphically by
phenograms. However, there is a serious problem of evaluation of
reliability and number of hierarchical levels recognized in the given
tree topology. The one decision might be based on analysis of
distribution of patrisitic distances, but again the problem of
probabilistic assessemens of non-zero distanmces remains which, in its
turn might be resolved by Monte Carlo simulations. However, to me
perslonally, bootstrapping pocedure looks more attractive as it allows
more straitforward bilogoical interpretation of numerical supports of
the tree branchs. I attempted myself to do that, and preliminary results
appear to be promising.

So, the question is, if there is anybody here knowning study cases in
which bootstrap is applied to phenograms and not to cladograms, and
correctness of such an application is considered. If someone knows these
cases, I would be extremely greatful for the references. It's needless
to say that any comments and ideas would be highly appreciated.

Yours sincerely,

Igor

- - -
Igor Ya. Pavlinov, DrS
Leading Researcher
Chief, Division of Mammals
Zoological Museum of Moscow M.V.Lomonosov University
ul. Bol'shaya Nikitskaya 6
125009 Moscow
Russia
E-mail [EMAIL PROTECTED]
Home page http://zmmu.msu.ru/personal/pavlinov/pavlinov_eng.htm
-- 
Replies will be sent to the list.
For more information visit http://www.morphometrics.org

Reply via email to