I'm reading On The Origin of the Species right now. He discusses human selection of domestic animals and plants at length as a fairly convincing analogy, then connects it to observations 'in nature'. It's hard to demonstrate something that is observed (vs something controllable), but the argument is quite comprehensive.