The concept of mimicry involves the idea that a species, for one reason or another, possesses certain characteristics (chemical defenses, for instance) that make it undesirable as a food item, and that other species (mimics) evolve a similar form to take advantage of the relative safety the undesirable species (the 'model') enjoys. There are two main types of mimicry: batesian and muellerian. These will be discussed as they apply to butterflies.
In batesian mimicry, the model is distasteful, and has aposematic coloration to advertise this fact to predators; the mimic, on the other hand, is quite palatable, but has also adapted similar coloration and form to resemble that of the model.
Predators learn to avoid a certain species through experience (Turner, 1987). Therefore when a batesian mimic becomes common enough, its advantage will decline; the chances that the edible mimic, not the inedible model, will be encountered increases. The fitness of the mimic therefore depends on the numbers of the mimetic form relative to the numbers of the model. If there are a number of batesian mimics of the same species coexisting in the same region, there will be a degree of polymorphism among them (Turner, 1987).
In batesian mimicry, there are other ways, besides coloration and form, that mimics rely on to ensure their safety. As their fitness depends on the illusion of inedibility, not only do there have to be more models than mimics, but the models have to be encountered first, in order to establish their inedibility; it is for this reason that most batesian mimics emerge later in the season than their models (Rothschild, 1971). They are also often smaller than their models (perhaps to decrease desirability), and have brighter versions of the model color pattern, which could warn the predator from far away to prevent it from taking a closer look and subsequently see through the mimic's ruse (Rothschild, 1971). Any mimic that is captured by a predator jeopardizes the fitness of the both the mimic and its host.
Muellerian mimicry is quite different from batesian. In this case, both the model and the mimic are distasteful; and while the fitness of model and mimic in a batesian system system decline as the mimic's numbers grow, an increase in the number of "mimics" in a muellerian system actually benefits both parties (Turner, 1987). Hence, different species of mimics of the same model coexisting together will have a developmental trend towards monomorphism of color, pattern and form (Turner, 1987).
How does mimicry evolve? In butterflies, color-pattern development is an active genetic process. The pattern of a butterfly's wing is composed of individual elements, such as bands, patches, and spots; each element is linked to one or more genes (Nijhout, 1994). In the best-studied mimicry relationship, the muellerian mimics Heliconius melpomene and H. erato, there are at least 39 genes that control development of their color patterns (Nijhout, 1994). Some genes control the size and presence (or absence) of certain elements, while others control patterns on only the forewing or hindwing, and still others may control related elements on both wings (Nijhout, 1994).
It has been suggested that batesian mimics inherit their color patterns as a single unit, a 'supergene' (Turner, 1987), and that very accurate mimicry has been acheived by a single macromutation, one large enough in its initial phase to confer at least some degree of advantage to the future mimic. How large must this initial mutation be? It rests entirely upon the ability of a predator to visually discriminate (Genoni, 1988). How well a bird can tell a distasteful model and its edible mimic apart dictates the level of detail necessary for the future mimic to survive and reproduce. In the batesian relationship between Battus philenor (pipevine swallowtail) and its mimic Limenitis astyanax (red-spotted purple), the fact that the model has tails on the hindwings doesn't seem to be a significant factor that the mimic needs to emulate in order to make its ruse work. Muellerian mimics do not depend on a supergene complex and macromutation; their protection lies in their distastefulness; it follows that they will also have at least some beginnings of aposematic coloration and pattern. This gives them a margin of safety that allows them to converge towards the model's form at a more gradual pace (Turner, 1987).
The concept of mimicry, if nothing else, demonstrates the remarkable adaptability of organisms to avoid predation. However, predators can adapt as well, and this is nicely exemplified by the feeding habits of the grasshopper mouse, Onychomys torridus, which feeds on various desert beetles including, those in the genus Eleodes (Tenebrionidae) (Smith, 1986). Eleodes beetles are equipped with a noxious chemical defense system that they advertise by doing "headstands," and exposing the abdominal glands that exude their chemical secretion; this secretion is composed of benzoquinones and alkenes that burn the eyes and mucus membranes (Hetz and Slobodichikoff, 1988). Onychomys has adopted a novel way of thwarting this defense. It grabs the beetle and plunges its abdomen into the sand, where the chemical is released harmlessly, and the mouse can eat the top portion of the beetle (Smith, 1986). There is a mimic of Eleodes beetles, another Tenebrionid, Stenomorpha marginata, that does not possess the chemical defense, but is similar in appearance and mimics the headstanding behavior (Hetz and Slobodichikoff, 1988). In the light of the mouse's feeding strategy, this mimicry is completely useless. While this situation between mimic and predator tends to be the exception rather than the rule, it puts a wrinkle into the security of btoh batesian and muellerian mimicry systems.