[phylocom-user] Treewise trait conservatism - comprehension

Scott Chamberlain schamber at rice.edu
Thu Nov 12 13:41:47 PST 2009


 Output from "aot" in phylocom gives the following columns of data for
treewise conservatism (from table 3):
*varcn *(variance of standardized contrasts across tree),
*varcn.ranklow*(signficiance of contrast variance relative to null
model, low rank), and
*varcn.rankhi *(signficiance of contrast variance relative to null model,
high rank)

For the two traits in my dataset, I got the following for mean height in the
order above (varcn= 8.08, low= 0, high=10000), and for maximum height
(varcn= 77.7, low=10, high=9990).

These are obviously not correlation coefficients as in Cavender-Bares et al.
2004 in which they were testing for trait conservatism. In Blomberg et al.
(page 719) they suggest low values of the variance of standardized contrasts
across a tree suggest that related species tend to share similar trait
values (or traits are "conserved"), whereas high values of variance suggest
"antisignal" which I assume is equal to "convergence".

So in my results above, mean height is relatively more conserved, and
maximum height is relatively more convergent? Or, both are conserved because
most randomization runs are above the observed values (10,000 and 9,990,
respectively)?

Thank! Scott Chamberlain
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.phylodiversity.net/pipermail/phylocom-user-phylodiversity.net/attachments/20091112/d3d289b1/attachment-0001.htm>


More information about the phylocom-user mailing list