A report in the journal Molecular Ecology (see here) discusses salamanders that didn't need to be isolated for speciation to occur. New Scientist reported it (here) somewhat breathlessly, saying that analysis suggested gene flow between cave and spring salamanders was still occurring for some time after the species diverged.
Cave salamanders, of course, demonstrate typical adaptions for an alternative environment, such as attenuated visual circuitry (good eyesight is burdensome to maintain), and heightened non-visual sensory systems.
I see nothing out of the ordinary in this finding. There's nothing to stop commingling even after the species adopt divergent characteristics (until genetic or behavioural differences sufficiently preclude). Simply put, as long as they can continue to breed, the offspring may or may not have the adaptive characteristics. Those that do, won't flourish in the new encironment, those that don't, won't.
It's not entirely that simple - that's a very broad approximation. In the period of interbreeding, there are a number of possible statistical outcomes.
But I'm not sure why New Scientist made such a meal of this. Any suggestions, disagreements welcome.