[Mesquitelist] Pairwise comparisons

[Michel Laurin]

Dear colleagues,

I am using Mesquite to perform pairwise comparisons, a test described by:

Maddison WP (2000) Testing character correlation using pairwise 
comparisons on a phylogeny. J theor Biol 202: 195-204

However, I have some trouble interpreting the output. Among the 
statistics are the following:

Selector: Most Pairs
180 pairs
Positive 1
Negative 1
Neutral 11
Remainder 167

I understand this as specifying the way pairs are selected (most pairs 
gives the highest number of pairs, no matter the state in the terminal 
taxa).

180 pairs are made (logical given that the tree has 361 terminal taxa).

However, when I try to follow the pairs, I find much more than 13 in 
which both taxa are scored for both characters. Hence, I am puzzled by 
the 167 "Reminder", which I thought were the pairs for which at least 
one of the characters was not documented in at least one of the two 
compared taxa. Is it possible that polymorphism also results in these 
pairs being lumped in "remainder"?

I assume that "Positive" means that both characters covary in the same 
direction; "Negative", that one increases when the other decreases, and 
"Neutral" is when at least one of the characters does not vary.

Any information on this would be helpful.

Best wishes,

Michel Laurin

-- 
UMR 7207
Muséum National d’Histoire Naturelle
Batiment de Géologie	
Case postale 48
43 rue Buffon
F-75231 Paris cedex 05
FRANCE
http://tolweb.org/notes/?note_id=3669