Your experiment is ridiculous. The problem is twofold. Moving the food into water isn't going to turn the flies into fish. What it's far more likely to do is force the flies to go after a different source of food. Those who depend on the food you've moved will simply die out slowly (you'd best be moving the food over a period of a few million years with no other changes to the environment or all the flies will simply die) while those who have an alternate source of food will simply ignore the food moved under the water. Your example also seems to point toward flies turning into whales. The problem there is that you've skipped probably a few tens of thousands of different species that would need to evolve over tens or hundreds of millions of years to get from fly to whale. You're also ignoring the fact that species do not respond exclusively to a single stimuli. Every aspect of the environment in which they live would have to be controlled, and an extreme amount of knowledge and foresight would be required to create an environment that yields the traits necessary to "push" the development of the proper traits at the proper time via natural selection.
It's a monumental task, and not likely one humans will ever manage to approach achieving in a lab or a field.
Your experiment is ridiculous. The problem is twofold. Moving the food into water isn't going to turn the flies into fish. What it's far more likely to do is force the flies to go after a different source of food. Those who depend on the food you've moved will simply die out slowly (you'd best be moving the food over a period of a few million years with no other changes to the environment or all the flies will simply die) while those who have an alternate source of food will simply ignore the food moved under the water. Apparently I didn't share my vision enough. In my experiment, the flies are contained in a large tank. The tank has plenty of water, air, food and whatever else is needed to make a wonderful fruitfly habitat. Also contained in the tank is water. The food is slowly moved into the water, over generations of fruit flies. Each step along the way would require the fruit fly to evolve a little more.
For example, food at the surface of the water would kill off the fruitflies who were stupid enough to land on the water. If the food were just below the surface, then fruitflies who could dunk their little fruit fly heads in the water and eat would have the advanatage.
Theoretically, after thousands of generations of fruit files, and if evolution of species is true, then eventually the fittest fruit flies will have evolved to be able to swim to the bottom of the tank and get the food.
It would be a graudual process for sure, but it should be easily demonstrated. I wonder if evolutionists have tried it yet. If not, why?