Doesn't look like it helps your case much. Comparing the two runs with the same total weight but one having it on the bike and the other with the weight in the tires they observed a difference of 27 sec. out of a total time of about 52 min. or 0.9%. But the run with the weight in the tires also had a lower average power input (275 W vs. 277 W), which amounts to a difference of 0.7%. Those two figures (.7% and .9%) are within the bounds of accuracy of the reported results, so the experiment is entirely consistent with the claim that during a continuous climb there is no difference in the speed whether added weight is in the rotating tires&wheels on in the fixed components of the bike and rider. [I'd also note that rotating weight would have a minor extra effect during the initial acceleration at the start of the climb, and I agree with cplager's concern that putting water inside the tire may affect rolling resistance.]

Also interesting comparing the normal bike to one with reduced tire pressure (runs 3 and 4). The run with reduced pressure (only 3 bar) was slower by 58 sec. or 1.9%, but this run was done with even lower average power, 273 W vs. 278, or 1.8% less. So the reduced pressure appears to also not affect the speed under these conditions.