In regard to reduce the influence of air bubbles inside the sand filter column, CO2 was added into the filter column. It was expected that the CO2 will be dissolved in water and the air bubbles will no longer disturb the flow of the water. Two columns that have the same configurations were tested for their flow behavior especially in their hydraulic loading rate. The filter columns were operated intermittently with variable supernatant level (decreasing level). As can be seen in the figure, those two columns have the different behavior although they have the same condition. Those two columns were operated during the day only and in the night they stay saturated without any operation. Every time the filter columns were started in operation on the next day, the hydraulic loading rate decreased and increased along with the number of filtration. Generally, the hydraulic loading rate decreased slightly along with the number of filtration. Does anyone have any experiences on this matter? How could we explain this behavior?