DATA:      van Wijngaarden (unpublished)
DIRECTORY: \CANOCO\SAMPLES\PERMUTIO\BACI3SIT
PROBLEM:   Testing the effect of a toxicant on a species assemblage 
           in a three-site BACI design
ILLUSTRATION OF:
+ How to interpret the fit-diagnostics of species in the solution file
+ Details of the Take and Skip rule when some samples are deleted
_______________________________________________________________________________
FILE TYPE    |  FILE NAME    |         DESCRIPTION
-------------+---------------+-------------------------------------------------
 SPECIES     | species.dta   | 59 species in three ponds sampled once in 
             |               |  consecutive months
ENVIRONMENT  | toxicant.dta  | treatment indicators of the pond (Control, Low 
             |               |  and High dose) and coded as 0,1,2 in the 
             |               |  variable Toxlevel
 COVARIABLES | pondtime.dta  | pond and time indicators
 PROJECTS    | bacirda1.con  | permutation test of "no impact" against a 
             |               |  "level of impact" model (linear in Toxlevel 
             |               |  using 1 degree of freedom)
             | bacirda2.con  | permutation test of "no impact" against an 
             |               |  "impact per dose dosage class" model (2 degrees
             |               |  of freedom)
             | bacirda3.con  | as bacirda1.con but after deleting a species 
             |               |  that had extreme influence on the analysis 
             |               |  (species 8, clo dip)
             | baci_c_h.con  | re-analysis of Example 8.3.8 by deleting 
             |               |  the data from the Low-impact pond
-------------+---------------+-------------------------------------------------