rationalskepticism.org

E. B. Ford conducted experiments on this way back in 1947, and demonstrated empirically that even incomplete mimics perform better than non mimics. I've presented the paper in the past. The moment some degree of mimicry of a suitable model exists, that will form the basis for further developments. One of the best examples of mimicry I know of centres upon a butterfly called Heliconius ismenius, which, as an adult, is an exact copy of a butterfly from a different clade, namely Melinaea messatis. However, interestingly enough, these two are not part of a Batesian mimicry coupling. A Batesian mimicry coupling arises when a species lacking defences mimics a species possessing defences, frequently in the form of painful stings or toxic flesh. In the case of the aforementioned butterflies, both species are chemically defended. They form an example of a Müllerian mimicry coupling, where species with defences converge to look like each other, in effect, adopting a common standard for aposematic colouration, thus giving insectivorous birds less work to do. If several defended species possess radically different colour patterns, the birds have to learn all the colour patterns in question before learning to avoid those species. If several defended species share a common basic colour pattern, the birds only have to learn from one bad encounter with one of those species, before getting the message and avoiding them all. Standardised aposematic colouration thus benefits, in the long term, both the prey (fewer of them fall victim to attrition by the inexperienced before the predators learn not to make that mistake) and the predators (the predators have to engage in less labour in order to avoid unpleasant morsels).