subroutine mtcalc(single)
      double precision double
      real single
      double = grnd()
      single =real(double)
      return
      end