Can artificial selection be used to create new species?

What degree of difference, in taxonomic terms, has been produced by artificial selection in domestic animals? There are two answers, depending on what criteria a species is defined:

• Artificial selection has not produced much evidence for new reproductive species: most domestic dogs are interfertile and belong to the same species in a reproductive sense.

• The answer is different for the species concept based upon phenotypic appearance. Take the variety of dogs (Canis familiaris). To most human observers, the difference between extreme forms, such as a pekinese and a labrador, is much greater than that between two species in nature, such as a wolf (pictured opposite) and a jackal, or even two species in different genera, such as a wolf and an African hunting dog (Lycaon pictus).

The evidence from domestic animals suggests that artificial selection can produce extensive change in phenotypic appearance - enough to produce new species and even new genera - but has not produced much evidence for new reproductive species.