The attachment may not have worked the way I thought.
Here it is as text.
Seattle, August 4, 2022
Dear MorphMetters,
At the beginning of April I wrote you to announce that a
new manuscript of mine, "Dimensions of Morphological Integration,"
was under review at
the journal Evolutionary Biology and also was posted for your
perusal on Research Square, the free preprint service
associated with Springer journals. The gently revised
version of this article accepted six weeks ago appeared
just today on the journal's website as an Open Access offering.
I commend it to your attention for several reasons,
including the two new diagrams it demonstrates: the
"spiderweb diagram," a version of
the thin-plate spline grid using polar coordinates instead of
Cartesian coordinates, and the "pseudoprocrustes plot,"
a novel application of the Procrustes algebra for
comparing the Boas mean form to the Boas first principal component.
The paper's discussion of morphological integration is set
as a major modification of the celebrated Olson and Miller
(1958) approach in order to accommodate today's image-based
computational GMM workflows.
I'd be particularly interested in any thoughts you may have about
two of the paper's core arguments: the demonstration that
an excess of semilandmark count over landmark count
seriously interferes with efforts to understand
integration by any of the standard GMM tools, and the suggestion
that evolutionary scientists avoid computing the
statistic called V_rel (a formula based in the variance of the
eigenvalues of a morphometric correlation matrix) whenever the
data come in the form of 2D or 3D coordinates --
V_rel discards far too much potentially important information.
I'm also suggesting that reports of GMM integration studies
change their language, eschewing questions like "Is this
data set integrated?" or "How integrated is this data set?"
in favor of the much more subtle query "How is this data
set integrated?" Finally, the paper reveals a hitherto-unnoticed
failure of the thin-plate spline to correctly represent deformations
incorporating directional substructures like sheets
or arches for which thickness is a more salient quantification
than position.
I'd welcome any comments you feel like posting (or sending
to me privately) about these or any other aspects of my
suggested reformulation of Olson and Miller for today's
GMM integration studies. Whether you agree or disagree,
either way, I hope you enjoy the paper.
Fred Bookstein
[email protected]
--
You received this message because you are subscribed to the Google Groups
"Morphmet" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To view this discussion on the web visit
https://groups.google.com/d/msgid/morphmet2/20220805005224.GA12891%40shape.localdomain.
