That gravity can keep planets in orbit is not quite the same as the conjecture that evolution can change the form of living things over many generations. Form, like what? Like a cat becoming a dog, or vice versa? What forms are you talking about?
Gravity is physics, evolution is biology, so that is one difference.
No, we do NOT expect a cat to become a dog. Or, for that matter, any modern species to become any other modern species.
What is expected is that species change over time to adapt to their environment. Occasionally, populations split and the subpopulations change in different ways. Over many generations, the small changes add up to give big changes. But the descendants will still have many of the features of the ancestors. They won't be 'totally new'. But they will be different.
Even many evolution deniers allow for changes inside of 'kinds'. So, for example, a change from an original type of cat to the variety we see today: houecats, tigers, lions, cougars, etc. We can see the family resemblance even today and yes, they all had a common ancestor.
Similarly, there are a wide variety of different types of bears. Black bears, polar bears, sun bears, etc. And we can all see the family resemblance between them (which is why we call them all bears). And I don't think it is too much to say they all had a common ancestor.
So, the common ancestor of the cats was a mammal, it was a carnivore (a type of mammal). And while it shared many characteristics with modern cats, it did not share ll of them (size, coloration, etc). It looked something like a merger between the cat species we had today.
Similarly for bears. The ancestor would have had characteristics of bears, but not all of them. It would have looked similar to a merger of the modern bears.
But what we can realize is that the original cat and the original bear weren't all that different from each other. And the same can be said for the original canine and the original versions of the other carnivores. And the similarity at that stage *looks* like a family resemblance of the type that cats have today (or bears, or canines).
And so we can look for a common ancestor of the mammalian carnivores. And guess what? We find such in the fossil record: something that has characteristics of all the carnivore families, but not identical to any of them. A species that has 'family similarity' to all the 'original modern families'. And that was the population that was the last common ancestor of all modern mammalian carnivores.
We can do the same for other families of mammals: the horse family (horses, donkeys, zebras, etc) are clearly related. I doubt you would have difficulty seeing they can have a common ancestor. And we can then look at other animal families and see how *those* families are related.
THIS is evolution: a change in species over time, with populations splitting to be come new species. All descendants have characteristics of the ancestor species, but they have also changed to be individual species. This happens in response the differences in the environment. And it produces a tree-like structure for the inheritance: an original carnivore splits into species that will be the original feline, ursines, and canines (as well as others). Those then split to gives the modern cats, bears, and dogs.
Many deniers of evolution, with their notion of 'kind' actually accept a large amount of evolution: between different types of cats, for example. They agree on the last few leaves of the trees, but fail to see that the ancestors were also related and had a common ancestor.