-------- Original Message --------
Subject: Re: Procrustes distance
Date: Mon, 3 Nov 2008 06:27:28 -0800 (PST)
From: Dennis E. Slice <[EMAIL PROTECTED]>
To: [email protected]
References: <[EMAIL PROTECTED]>
Drat. I would rather have had the sentence in my last post read
something like:
...reduces the about of detectable "true" shape difference...
The point being, of course, one can only detect shape difference
represented in the set of landmarks analyzed, and morphologically
diverse specimens may limit the shared landmarks available for analysis.
-ds
morphmet wrote:
-------- Original Message --------
Subject: Re: Procrustes distance
Date: Mon, 3 Nov 2008 06:01:12 -0800 (PST)
From: Dennis E. Slice <[EMAIL PROTECTED]>
To: [email protected]
References: <[EMAIL PROTECTED]>
The concept of "big" depends, of course, on the within-sample
variability, hence the need for statistical tests.
That said, Procrustes distance is constrained to be <=pi/2 or about 1.57
radians (90 degrees) or sqrt(2) as a cord distance. A Procrustes
distance of 0.031 is about pi/100 (~1.8 degrees) whether in radians or
cord distance.
The largest zoologically meaningful Procrustes distance of which I am
aware was reported by Marcus et al. (2000) who compared the orders of
living mammals using cranial landmarks. The dolphin-to-muskrat
comparison produced a Procrustes distance of 0.731 radians (~42 degrees,
or ~pi/4). But note, the constraints of finding common landmarks on such
taxonomically disparate organisms likely reduces the amount of true
"shape" difference - whatever that means. This was recognized by Marcus
and colleagues.
-ds
Markus, L. F., E. Hingst-Zaher, H. Zaher. 2000. Application of landmark
morphometrics to skulls representing the orders of living mammals.
Hystrix 11(1): 27-47.
morphmet wrote:
-------- Original Message --------
Subject: Procrustes distance
Date: Mon, 3 Nov 2008 04:32:04 -0800 (PST)
From: Rebeca <[EMAIL PROTECTED]>
To: <[email protected]>
Dear Morphometricians,
I am studying shape variation in four populations of fish coming from
different geographical areas using a landmark- based GM approach.
I am doing permutation tests to check for mean shape differences between
the populations using the Procrustes distances, and a question about
these distances came up: Can we tell if a Proc. distance is fairly big
or small by seeing just the numbers? For example, if the distance
between two mean shapes is 0.031, is this a fairly big difference or not?
My question is more like when we describe the variation in the data, we
have a mean and a standard deviation, if the mean is 34.5 cm, and the sd
is 12 cm, I would say, by looking at the numbers that the sd is quite
‘big’.
Considering I am not an expert in morphometrics nor statistics, I hope
someone can help me to understand better this issue!
Thanks in advance,
Rebeca
Rebeca P. Rodriguez Mendoza
Email: [EMAIL PROTECTED]
Marine Research Institute CSIC
– Fisheries Dept.
C/ Eduardo Cabello, 6
Tel. 986 23 19 30 ext. 254
Fax: 986 29 27 62
36 208 Vigo, ESPAÑA
--
Dennis E. Slice
Associate Professor
Dept. of Scientific Computing
Florida State University
Dirac Science Library
Tallahassee, FL 32306-4120
-
Guest Professor
Department of Anthropology
University of Vienna
========================================================
--
Replies will be sent to the list.
For more information visit http://www.morphometrics.org
