I believe that Chihuahaus and Great Danes are reproductively isolated because of physical size differences and not by genetics.
The most commonly used, and most useful, species definition is (basically) reproductively isolated groups. It doesn't matter what is going on at the genetic level, if mate recognition, behavioral differences, or physical differences prevent gene flow between two groups.
Many cricket species are genetically compatible but they don't recognize each other's mating calls; some drosophilia species are genetically compatible but their genitalia can't interact due to morphology. An astounding example is the cichlid flock of Lake Malawi, a thousand or so species that are essentially all genetically compatible with each other, yet rarely interbreed in the wild. It is essential to include the non-genetic element in a species definition because otherwise we would discard all of the these isolated gene pools. A two-foot-long barracuda-like cichlid and a two-inch-long clownfish-like cichlid would be considered the same species because a hybrid could be produced in a petri dish, even though if they ever met the latter would immediately become the former's lunch.
All of that said to say this: Great danes and chihuahuas are probably still the same species, but in the way that the two extreme ends of a ring species are. Though they likely would never produce offspring on their own, gene flow is readily accomplished between intermediate-sized breeds (maybe only one might even be necessary).
Thus the genes of the great dane and the chihuahua can intermingle fairly easily without any rattling of test tubes...
[AbE: If the world decided to undertake the abhorrent experiment of killing off all non-chihuahua/non-great-dane canis in the world, we'd probably have ourselves two separate species...]
This message has been edited by pink sasquatch, 12-07-2005 06:36 PM