I'd say there's as much loss of precision as there is in pouring from one container to the other (there's always left over water in the original container...even if it's just drops)...
I fully admit that my method has a number of built in assumptions:
My original assumption that the beakers are perfect cylinders/rectangular prisms
The containers hold EXACTLY the number of gallons represented...not a drop more, nor less...(of course, this causes problems in that the second you MOVE a container to pour into another, you would lose some...unless you had an infinitely steady hand...)
You h