	---------------------------------------------
	Changes from version 2.0.12 to version 3.0.0
	---------------------------------------------

******
* R
******

- Function Fstat
	 . reformat genotypes first

- Function Fstat.output
	. extra arguments: coord and  burnin

- Function mcmcFmodel becomes function MCMC
	. test ncol(coordinates) only if passed as an argument
	. test freq.model=="Correlated" & varnpop==TRUE 
	. test nb.nuclei.max > 3*rate.max
	. init nb.nuclei.max depends on whether spatial is TRUE
	. 3 bugs corrected in bdpop9
	. new function sm: same job as (corrected) bdpop9 but faster
	. requires prior parameters shape1 and shape2 as input 
	. new variable output.files specifying output files to be created
	. possible values "Falush" and "Dirichlet" for parameter freq.model become "Correlated" and "Uncorrelated"
	. freq.model set to "Correlated" by default
	. npopmin set to 1 by default
	. npopinit set to npopmax if missing

- Function simFmodel
	. remove argument seed

- Function PosterioMode
	. plots coordinates
	. correction of a minor bug that created inconcistency between modal pop maps of indiv and pixels 
	in case different pop have same posterior proba

- New function PlanarProj
	. project points with coordinates given as (lon,lat) on a plan and return cartesian coordinates
	

- New function PostProcMultChain
	. concatenate the output of sevral MCMC runs  
	into a single run and postprocess it 


	
*************
* In Fortran
*************
	
- All computations in double precision	

- Use R RNG instead of randlib	
	
- mcmcfmodel become subroutine mcmcgld 
	. allele freq initialized form the full conditionnal
	. new variable ofile specifying output files to be created
	
- PostProcesschain 
	. attempt to relabel populations to avoid label switching
	. have two extra arguments regarding I/O 

- Subroutine udcf 
	. do not compute MH ratio if ipop1 .eq. ipop2
	. more direct computation of ratio (does not depend on freq)
	  + error in parenthesis in algama

- New subroutine sm and smd (replace bdpop9 and bdpop9bis)
	. attempt to split and merge populations
	. computes ratio more directly (not depending of frequencies)
	. attemt to split a pop even if 0 of its cells are selected
	. corrects errors in ratio on split-merge with Correlated model: 
	     . q=1-d/d instead of d/(1-d)
	     . normalising constant wrongly coded
	. propose new drifts through a RJ technique instead of from the prior    
	
- Subroutines udcf, udcf2 and updurw	
	. attempt to update the color or location of one tile only 


- New function postprocmultchain


- New function pppmindivmultchain