To simulate the introduction of new mutations, we simply
add a random integer to the new child. Thus to create a new child
named ` x` from the ith member of the current generation, the
code is:

x = g[i] + poisson(u);where

Note that the variable ` u` corresponds to the parameter
of the mathematical model.
For small values of the expected value the Poisson procedure
described in * Numerical Recipes* [3]
works fine. The average
number of mutations per individual measured empirically over a wide
variety of species is about 1.0, and the procedure in * Numerical
Recipes* works for , so it should be sufficient.