[R-sig-eco] Fit of ordination axes always given?

[Jens Oldeland]

Dear group,

I have used the as.mlm.cca(vegan) function to derive statistical 
measures for the goodness of fit of my environmental variables and the 
ordination axes of an RDA.
I prepared a manuscript in which I used spectral indices (NDVI etc.) as 
environmental variables in a RDAand have now received answers from a 
reviewer in which he/she states that:


/'Ordination axes that show a good statistical fit with the spectral 
indices' - all ordination axes are linearly related to the predictors - 
this is forced by the CCA or RDA method itself (see l. 203 ff). Thus, 
statistical fit is always given. /


I agree that all ordination axes are linearly related to the predictors, 
but I refuse to believe that therefore a good (!) 'statistical fit is 
always given'. It might be also pretty poor, in my opinion.

Am I wrong?


I would be very grateful if someone could comment on that,
thank you in advance,
Jens
//

-- 
+++++++++++++++++++++++++++++++++++++++++
Dipl.Biol. Jens Oldeland
University of Hamburg
Biocentre Klein Flottbek and Botanical Garden
Ohnhorststr. 18
22609 Hamburg,
Germany

Tel:    0049-(0)40-42816-407
Fax:    0049-(0)40-42816-543
Mail: 	Oldeland at botanik.uni-hamburg.de
        Jens.Oldeland at DLR.de  (for attachments > 2mb!!)
http://www.biologie.uni-hamburg.de/bzf/fbda005/fbda005.htm
+++++++++++++++++++++++++++++++++++++++++