runtime error 206: floating point underflow

How can I evade runtime error 206? I want the program to substitute a zero when
necessary and have read in the Borland manuals that this is possible, but they
told not how.. Who knows how to do it?

Greetings,

Martijn Brouwer                 e.a.m.brouwer@No$pam.student.utwente.nl

                                remove No$pam.