The hallmark of life on earth is morphological diversity,
which is represented in the spectacular sexually dimorphic and
polymorphic wing patterns of butterflies. Biologists have long tried to
find general ecological explanations for the evolution of morphological
diversity, especially at intraspecific levels. We address this problem
by investigating the remarkable diversity of mimetic wing patterns in
butterflies, manifested in numerous sex-limited or otherwise polymorphic
forms within species. Historically, the major hypotheses to explain
these polymorphic forms have centered around the concepts of sexual
selection, physiological tradeoffs, and shifting balance. While there is
interesting evidence to support components of each of these ideas, a
broad-ranging ecological explanation for this type of polymorphic wing
patterns is still missing. Towards this end, we present a simple
mathematical model that may explain the full gamut of polymorphic
diversity exhibited by mimetic butterflies. Our model is based on
frequency dependent natural selection (predation) and sex-specific
advantages of mimicry, both of which are empirically established. Our
model predicts that in ecological communities composed of toxic models
and undefended mimics, different threshold mimic frequencies favor
transitions between monomorphic, female-limited and polymoprhic mimicry.
Data on 57 mimetic butterfly species showed that monomorphic mimicry was
prevalent at low mimic frequencies whereas female-limited and
polymorphic mimicry was prevalent at high mimic frequency, as predicted
by the model. We further show with controlled experiments on the Papilio
polytes butterfly that female-limited mimetic polymorphism in this
species could not be explained by sexual selection, physiological
tradeoffs or other hypotheses. However, the presence and density of
models appeared to determine the presence and frequency of mimetic forms
in a given locality. Thus, frequency-dependent natural selection in the
form of predation appears to be the predominant force driving wing
pattern polymorphism in Batesian mimetic butterflies. We further
elucidate the genetic mechanism underlying this polymorphism in Papilio
polytes. We show that double sex, a transcription factor that controls
sexual differentiation in insects, also controls sex-limited mimetic
polymorphism via sex- and tissue-specific alternative RNA isoforms and
differential expression. Polymorphic wing patterns in other species have
also been shown to be controlled by master gene regulators that have
different functions in other insects. In all these cases, the master
regulators have been co-opted from the original function to produce
polymorphic wings in butterflies. Our results suggest that the genetics
of highly complex and diverse phenotypes may be relatively simple, and
may involve frequent exaptations.