My ELISA plate has 12 replicates of each sample in a row, and it shows an OD gradient. for example A1 (100mIU) if say has OD of 3.9 and gradient down to 2.9 OD at A12. I addressed different parameters and tightened the CVs, but gradient still persists. My background and negative controls are 0.07 and the gradient is mild when compared to my 100mIU samples. Have any of you seen this? what could cause this and any suggestions to improve this?