The measurements I made track fairly well. There is a bit of variation, expecially from the DXB numbers which are sparse since it is only given in 2 inch increments: 2,4,6; 4 data points compared to TTaylors 12 data points and my 19 data points.
At 2 inches TTaylor and I match at 13 gallons, but then at 4 inches we differ by more than a gallon. That trend of differing a gallon or more continues until 24 gallons at 6 inches, where all three sources converge again.
I will probably do the experiment again. It's tough to drain those tanks. I removed the drain plug to drain the tank and I probably should have had my buddy with me to help because that gas comes out very fast and fills 5 gallon cans really quick. After I filled the 5 gallon tanks, I used a 1 gallon can to measure each gallon and fill the tank. I waited about 2 - 5 minutes in between to give the tank time to settle before measuring.
I started with the gas barely covering the bottom of the visible rubber bladder tank. Assuming that is 6 gallons could be off by a bit, so I'd really like to start with 0. Then you need to consider how much of that gas is unusable.
I looked through the engine log, the airframe log and the weight and balance sheets but I can't find a mention of unusable fuel with the install in 2002 of the ON bladder tanks. I have heard on this site of a mention of 2.X gallons of unusable fuel in the bladder.