How can they tell it was a different species from Lucy just by a jaw bone. By species, I mean not able to inter-breed; unless I have the definition of species wrong.
Species is a useful construct inside one era: This group is not reproductively compatible enough with this other group that they blend together when the two populations come into contact.
It's not a particularly useful construct over time - I can't say how far back I would have been reproductively compatible with my ancestors, and all internally-compatible population groups drift in phenotype over time and geography. We would need many orders of magnitude more fossil evidence to begin to form useful natural categories by clustering primary phenotypical characteristics, this far back. The picture is a lot clearer for the Holocene extinctions, particularly Arctic and cave-dwelling species, where we don't have to rely on the rare long-lived fossil... but even with present-day extant populations, a lot of the predictions taxonomists made about relationships before genetics offered extra objective data, have proved incorrect.
The hypotheses put forth for the human family tree, given how little evidence we have, are barely science - untestable guesses that simply try to connect the dots between all presently-available fossil fragments.
The pre-avian dinosaurs were around for ~165 million years and we only have a few thousand complete skeletons to cover that entire span. Not only is it basically impossible to divine species there using statistically valid distinctions, genus is right out too - most complete skeletons will be the only complete representative for the whole family, and the error rate on classifying small numbers of bone fragments is such that low-completeness specimens are statistically useless for drawing lines. It's like performing clustering analysis at n=1 or n=3 or n= (2 + 0.02 + 0.01 + 0.13)
Paleontology draws lines anyway, but it's largely an academic exercise in guesswork: How do you tell which species in the ~hundred-species family this thighbone came from? If it's not exactly the same size or found at the same site-stratum as the type specimen for the family, or there are any notable shape differences, just make up a new name for it, the second known genus in the family.
It's a bit more complicated than that. The edges of species can be very fuzzy. For example, it's possible that population A can interbreed with population B, which can interbreed with population C, but A and C cannot interbreed.
It's also possible that populations can interbreed, but don't, due to differences in behaviour, looks or environment.
Are dogs and wolves the same species? They can interbreed, yet are seen as different species. What about Great Danes and Chihuahuas?
Humans and Neanderthals interbred, we know now, and yet are different species.
