Species is most easily defined in extant sexual populations, since you can observe reproductive compatibility. Two populations that cannot or will not breed are considered separate species. Just in the last few months I've noticed that most creationists have punted on claiming that this does not occur, quietly moving the goalpost for macroevolution to the genus level.
In asexual species, you have molecular and anatomical evidence, but it's not so clear as for sexuals. For fossil species, all you have to go on are anatomical clues. That's fuzzier yet. You can picture future paleontologists arguing over St. Bernard and chihuahua skeletons.