It seems that many of the field labs had performed gradation tests with worn sieves. While most differences were minor, they introduced a degree of unnecessary variability when combined. How can we can minimize our testing variables on these large projects?