I recently replied to a question about minimizing variability in the volume of liquid present in a well at the end of a cell-based RNAi experiment. It is important to minimize variability in the volumes well-to-well, as variability can affect the assay and/or the assay readout, causing 'noise' in the screen that obscures meaningful results. Because others might have similar concerns, I'll post a summary of my answers here.
Variability in micro-plate well volumes is usually due to equipment issues or evaporation. We recommend the following steps to limiting problems.
(1) Have your pipetting or automated liquid handling equipment calibrated, and keep it clean and clog-free. If one row or column is consistently different from the others, that can suggest you have a clog or another equipment problem.
(2) Make sure the equipment you are using is accurate in the range of volumes you are using. You would not use a P-1000 hand-held pipettor to measure 2 microliters of volume, as a P-1000 is not considered accurate in that range. Similar to this, automated liquid handling instruments are considered accurate in only within specific ranges, and should be used accordingly.
(3) In addition to using a humidified incubator, we recommend also putting the plates inside plastic containers with damp paper towels at the bottom, to help minimize evaporation loss.
(4) Take all appropriate steps to limit "false discovery" due to any cause, including: establish good positive and negative controls, make sure you have a good signal-to-noise before screening, perform replicate tests when you screen, use appropriate statistical analyses to identify screen hits, and experimentally test initial screen hits with other reagents and assays.