I understand that this sort of thing is based upon homology (structure) rather than analogy, but somehow I think they mix things up a bit.
Correct; the idea is to look at shared traits which arose from the common ancestor. The trick is how you tell what's similar because of common origin, and what's similar due to convergence (this applies to both morphology, where convergent evolution complicates things, and genetics, where it's entirely possible for mutations to be 'reversed' by subsequent mutations).
I'm also not fond of how they determine speciation. It kinda goes against that old thing about phylogeny recapitulating ontogeny.
The idea is old and, specifically, disused. While it's poetic and *does* allude to the evolutionary significance of shared developmental stages, it's long since been recognized as wrong. Sadly, this recognition led to the separation of developmental biology from evolutionary biology until recently, when we're recombining the two with much more nuanced views.
How can they call things a species when the crosses produce viable offspring that can further produce viable offspring?
A quote I once heard went something like this: "If you want to see the most vicious blood-sport ever, lock a dozen biologists in a room and tell them to define 'species'."
Since all species are constantly evolving, it's very hard to draw a static box around them, especially when you take asexual species into account. Is every asexual gecko individual a new species, since they never interbreed? What about bacteria, who scavenge DNA from the environment and incorporate it into their own genomes? The "separate gene pools" idea is good in theory, but falls apart when applied to the diversity of the natural world, leading to lots of hand-waving, workarounds, and general confusion.
As for the particular issue of hybridization, what you're missing is post-zygotic barriers, which you rarely see in cultivation. What if the hybrids can't survive natural winters, or lack a crucial chemical to protect against a local bug? What if they've got weakend immunity, or aren't desirable to the opposite sex (or pollinators)? Numerous factors can create a barrier to gene flow in natural settings, even if it's just different polination/breeding times; real plants can't rely on a person with a fridge and paintbrush to store their pollen a few months.
Numerous natural species produce hybrids, and I've even personally seen (almost assuredly non-fertile) cross-
genus hybrids (corn snake x king snake and pygmy rattler x western diamondback). Darwin's finches, the flagship species of evolution, are known to hybridize during drought conditions (though such conditions are infrequent enough that the species will remain distinct). One species of frog (and several fish) exist only as hybrids, and can only reproduce by crossing with one of the two parent species (google hybridogenesis). Plants are even worse, as they hybridize far more readily than animals.
So basically, don't think in terms of absolutes; a trickle of gene flow doesn't undermine species status, nor does artificial hybridization.
Mokele