One of the many fascinating aspects of evolution is how nature happens upon the same solutions to environmental problems, over and over.
Two of the best examples of this are eyes and wings. True eyes evolved from simple photoreceptive spots at least four times: the vertebrate eye, with its complex system of lenses and retinas; the pinhole-camera eyes of the chambered nautilus and other cephalopods; the compound eyes of insects; and the rows of separate spherical eyes in clams and scallops. Wings, on the other hand, evolved independently no fewer than six times: bats, birds, insects, pterosaurs, flying squirrels, and colugos (the last two count if you include gliding along with true powered flight).
The reason is simple. There are a handful of problems animals have to overcome (perception/sensation, nutrition, reproduction, locomotion, avoiding environmental dangers, and avoiding predation) and a limited number of ways to accomplish them. Once (for example) photoreceptive eyespots develop in an animal, natural selection for improving the sensitivity of those spots takes over, but how exactly you do that can differ. The result is you end up with vision evolving over and over, and each time, the organ is structured differently, but accomplishes the same thing.
Evolution, it seems, is the law of whatever works.
This has interesting implications about what extraterrestrial life might look like. I very much believe that certain features will turn out to be constrained in any conceivable species -- the presence of locomotor organs, organs sensitive to sound, light, heat, and touch, and so on -- but also, that the way those organs are arranged and configured could be very differently from anything we have on Earth.
This "multiple solutions to the same problems" idea is what immediately came to mind when my friend and fellow writer Gil Miller, whose inquisitive mind and insatiable curiosity have provided me with many a topic here at Skeptophilia, sent me a link from Phys.org about hollow bones in dinosaurs. Endoskeletons such as our own exist in an interesting tension. They have to be solid enough to support our weight, but the better they are at weight-bearing, the heavier they themselves are. The mass of an animal in general increases much faster than its linear dimensions do; double a mouse's height, keeping its other proportions the same, and it will weigh about eight times as much. This is why in order for the whole system to work, the proportions have to change as species increase in size. A mouse's little matchstick legs would never work if you scaled it up to be as big as a dog; at the extreme end, consider the diameter of an elephant's legs in relation to its size. Anything narrower simply wouldn't support its weight.
[Nota bene: this is why if you were traumatized when young by bad black-and-white horror movies about enormous insects wreaking havoc, you have nothing to worry about. If you took, for example, an ant, and made it three meters long, its proportionally tiny little legs would never be able to lift it. The worst it could wreak would be to lie there on the ground, helpless, rather than eating Tokyo, which is what the horror movie monsters always did. One got the impression the inhabitants of Tokyo spent ten percent of their time working, relaxing, and raising families, and the other ninety percent being messily devoured by giant radioactive bugs.]
But back to the Phys.org article. A detailed analysis of the bone structure of three different dinosaur lineages -- ornithischians, sauropodomorphs, and herrerasaurids -- found that while all three had landed on the idea of internal air sacs as a way of reducing the mass of their large bones, the structures of each are different enough to suggest all three evolved the feature independently. Once again, we have an analogous situation to eyes and wings; identical problem, parallel solutions. The problem here is that large body size requires heavy bones that require a lot of energy to move around, and the solution is to lighten those bones by hollowing them out (while leaving the interstices connected enough that they're still structurally sound). And three different clades of dinosaurs each happened upon slightly different ways to do this.