Is there a better way to evaluate yield response?<BR>Frequently herbicide tolerance trials are NS <BR> <BR>Large LSD values?<BR>