I would like to have a function float_to_round (expr) which parses the expression, and for each element z, if floatnump(z) is true, converts that element to round(z), and returns the expression. example: float_to_round (bessel_j(3.0,x)*bessel_y(2.0,x)) ---> bessel_j(3,x)*bessel_y(2,x) Ted Woollett