Would it be more accurate to test the running time on an integration such as fint(x^(1/2),x,0,50)? --------------------------------------------------------- Get Your *Web-Based* Free Email at http://www.hotmail.com ---------------------------------------------------------