Much like supporting multiple populations, support of
multiple environments may encourage multiple solutions.
 And cross breeding between the different environment
may push a population onto a track of a more ideal
solution.

Benefits: (speculation)
    - Multiple solutions
    - Cross breeding of populations between
environments may provide the necessary bump out of a
local optimum.
    - Multiple environments allow definition of
multiple fitness schema.  Sometimes, fitness is hard to
determine.    It may be a compilation of many different
attributes.  This would allow for different
environments to select for different fitness criteria,
and the eventual merging of populations which may
provide an optimal solution relative to multiple
fitness determinants.
    - Multiple environments could support different
mutation rates and algorithms.  A particular rate or
frequency of mutation in one population may provide a
different solution than that of another environment. 
More importantly, the cross breeding of populations
between environments will effectively allow the
breeding of a population that has evolved over 100
generations with one over 20 generations, thus possibly
injecting information into the "more sophisticated"
population which may have been lost after so many
generations.