One of the pieces of nonsense creationists love to peddle is that evolution is not like gravity because it makes no experimentally falsifiable predictions.
Usually, when evolutionists try to combat the “only a theory” argument, they drag gravity into it. They say Newton’s theory of gravity is “just a theory, too.” The difference between gravity and evolution, of course, is that one can do repeatable experiments to test the theory of gravity. Engineers can measure the amount of force it takes to stretch a spring a certain length. Then, they hang various masses from the spring and measure how far it stretches. From this they can determine the force of gravity pulling on the mass.
Furthermore, the theory of gravity made some interesting predictions. Astronomers noted that some of the outer planets did not orbit in the path one would expect. They calculated that some other gravitation force must be acting on them. From that they calculated where an unknown planet must be. They looked in that location and discovered Pluto. The theory of gravity predicted a planet of a particular size in a particular orbit, and it turned out to be a correct prediction.
Now, if you read science blogs like Pharyngula and The Panda’s Thumb (and if you don’t, you really should), you might find yourself asking, “Isn’t evolution exactly like that?”. After all, evolutionary biology’s ability to predict the Chicxulub impact event was pretty amazing, to say nothing of some of the sequences of simple-to-complex animals found in the fossil record.
That’s not experimental, of course. Which, I suppose, is how some creationists excuse their “Evolution is anti-scientific” blather. It’s possible to do lab experiments in particle physics, or organic chemistry, or metallurgy, but not geology, evolutionary biology, or astronomy. The most you can do in evolutionary biology is make fruit flies speciate in the lab; the more interesting science comes from the fossil record or from genetic comparisons.
In observational sciences like the last three, falsifiable predictions are based on future observations: a radical change in the fossil record will correlate to evidence of global climate change or an impact event, a star that wobbles inexplicably probably has a planetary system, ice core samples indicating low temperatures in the last few million years will also indicate lower CO2 levels.
Personally, the evolutionary prediction I find the most breathtaking is the evolution of the eye. Darwin hypothesized how it might’ve evolved from simple light-sensitive cells. For someone working only from extant fauna, his predictions were extremely accurate.
But politically, one of the strongest points of evolutionary theory is the arguments Michael Behe et al field against it. Each time, they say that feature X couldn’t have evolved, in effect forcing a falsifiable prediction related to the fossil record on scientists. And each time, scientists manage to pass the test – the evolution of vision and color vision is well-understood now, paleontologists have managed to get a pretty solid grip on cetaceans’ return to the sea, and now biochemists are in the process of showing how the flagellum is reducibly complex.