Jeremy noticed that the PVC motor solution (orginally intended to be slung under the bottle) would fit inside the mouth of the bottle.

In this case, side floats were required to keep the boat from rotating. This worked (like in the first side-float design posted earlier) since the PVC weighed down the rear of the boat and kept the propeller submerged. The high angle of attack also let the bow of the boat ride up on the water and succesfully break the beam sensors on both ends of the track.