ver2.0.3
Modifications to c++ code to satisfy clang 3.4 issues.
ver 2.0.2
directly call fround rather than rround
ver 2.0
removed the dependencies on pegas, because it was causing segfaults on 64bit linux machines.  I don't understand the mechanism 
for these errors, but isolated to this package.  May create an rmetasim-analysis package in the future to sidestep this issue

Also increased speed for reproduction by better storing intermediate values.  In the case of complete selfing, speed is much increased

ver 1.1.15
fixed some gcc specific constructs and repaired some documentation
ver 1.1.14
fixed a lot of memory management issues
ver 1.1.13
backed out the generalization of memory management implemented in 1.1.10 which caused an ugly memory leak.  In the process of 
improving memory management.

ver 1.1.12
increased the length of DNA seqs from 500 to 1000

ver 1.1.11
1/23/10
Resolved a small problem with the Rd file for landscape.mismatchdist.

ver 1.1.10
1/22/10
Fixed gnu-specific C++ array allocation.  Should now compile on Sun.
Also added a function 'landscape.mig.matrix' that automatically creates a migration
matrix among populations

ver 1.1.09
10/7/09
Fixed changed dependency on ape and added 'pegas' functions as well as 
explicit call to ade4's amova function to distinguish it from the pegas version.

ver 1.1.08
10/4/09
Fixed problem assigning individual ids (Thanks to Karen Martien)

ver 1.1.07
3/4/09
repaired bug in landscape.sample()

ver 1.1.05
1/24/08
added code to interface with amova functions in ade4. 

ver 1.1.02
11/22/07
removed landscape.new.epoch.island and landscape.new.epoch.step
They didn't work as advertised and there has been limited desire to 
use them

ver 1.1.009
5/10/07

Reinstated the labelling of individuals, their mothers, and fathers in
the columns 4:6 in the individual sub-object of a landscape.
 
ver 1.1.006
12/16/06

Finally fixed "the bug" that Karen Martien pointed out last year.
This was very hard to find because it did not manifest on my
development platform (mac), nor in the way that I use the programs and
therefore the way that my testscript works.

ver 1.1.0
12/3/06 Fixed some stack imbalance problems.  Updated the code to read in simcoal simulations

10/10/06 Eliminated out of date documentation.  Need to develop a
Vignette for the package.  This is High on the TODO list

9/29/06 Tried to make rmetasim function names more consistent by
appending 'lansdscape' to the front of them.  This should reduce
naming collision problems with base and other packages in R

3/25/06 Fixed bug in conversion between rmetasim objects and internal
c++ representations.  Added code to parse Arlequin files into rmetasim
objects

3/25/06 IMPORTANT!!! Added three columns to individuals object between
the demographic information and the genetic information.  They store
an individuals unique ID, it's mothers unique ID and fathers unique
ID.  This can be used to track family structure over multiple
generations.  Right now, rmetasim does nothing but create and assign
these IDs properly.  This change alters the text file structure of
rmetasim and will have dire repurcussions for anyone who directly
accesses the individual objects and does not account for the addtional
three columns.  All of the rmetasim functions should work correctly.

If individuals are created outside of the c++ backend, they should be
given appropriate IDs Zero for mother and father indicates individuals
created de novo

Several new functions have been defined to work with these objects.  




ver 1.0.5
11/16/05  To accomodate changes in R 2.2 that implemented the S3 base
'simulate', changed the central function simulate.landscape() to sim.landscape()

ver 1.0.2
1/11/04   Fixed problem with is.landscape() AES
1/5/05    Addded some update features to landscapes AES
