DATA:      Aitchison (1984a: the Coxite data)
DIRECTORY: \CANOCO\SAMPLES\METHODS\LOGRATIO
PROBLEM:   How does the chemical composition of rock relate to depth and 
           its porosity?
ILLUSTRATION OF:
+ how to obtain a log-ratio analysis in CANOCO using PCA and RDA
+ how to regress composition on explanatory variables
+ problems of log(Ay+B) -transformation with B = 0
_______________________________________________________________________________
FILE TYPE    |  FILE NAME    |         DESCRIPTION
-------------+---------------+-------------------------------------------------
 SPECIES     | coxite.dta    | chemical composition of rock (5 chemical species
             |               |  in 25 samples)
ENVIRONMENT  | depthpor.dta  | depth and porosity of the samples
 DERIVED     | coxite2.dta   | percentages changed to fractions except for 
             |               |  samples S1 and S2
             | cox_zero.dta  | species C in sample S1 set to 0.00 in 
             |               |  an all-fractions file
 PROJECTS    | pca_lgrt.con  | Log-ratio PCA of compositional data 
             |               |  (log-contrasts)
             | pca_chck.con  | Log-ratio PCA on coxite2.dta
             | rda_lgrt.con  | Log-ratio RDA of compositional data using 
             |               |  log-ratios (with permutation test)
             | rda_forw.con  | manual forward selection of variables with
             |               |  permutation tests that account for the spatial
             |               |  arrangement of the samples
             | error.con     | Log-ratio PCA on cox_zero.dta with log(y)
             |               |  transformation
             | corr_err.con  | Log-ratio PCA on cox_zero.dta with log(100y+1)
             |               |  transformation
-------------+---------------+-------------------------------------------------

Aitchison, J.(1984a): The statistical analysis of geochemical compositions. 
  Mathematical Geology, 16: 531-564.