Think about how evolution fills niches in different parts of the earth. In Australia there are marsupials (a very primitive form of mammal that have evolved along a different evolutionary path for more than 65 million years since leaving South America before the continents drifted apart). But they evolved along fairly similar lines as mammals in certain respects according to the ecological niche they fill. They're different from mammals from other continents, but they're also very similar in many ways. Function does indeed seem to dictate form to a very large extent.

Good point Carol. Convergent evolution has produced strikingly similar forms filling given niches in widely divergent places and times. The plesiosaurs compare with cetaceans and for that matter sharks, rhinos resemble triceratops, and so on. But really all these are starting from a common body plan- vertebrates with a limb at each corner- that was set as long ago as the development of the chordates, and is really a contingency of our planet.

Certainly aerodynamic and hydrodynamic shapes will necessarily be moulded by the physics, but the moulding will be done on the basis of whatever basic foundation has evolved. There's nothing magical or even optimal about 2 legs and 2 arms, and arthropods find 6, 8, or many limbs work well. I could do with extra arms now and then...

And there's nothing necessary about bilateral symmetry, look at starfish, though having developed it has certain advantages in retrospect.

There may be other, secondary, considerations, like the size limits (in our environment) on arthropods, due to mechanical considerations and the need to grow by shedding the exoskeleton at intervals.

By the way, there's nothing "primitive" about marsupials.