We know that eyespots have evolved multiple times. This is because we observe the eyespots in several species with independent evolutionary histories. Put simply, the caterpillars of several butterfly and moth species have evolved the same solution for preventing or warding off attacks from hungry predators (mostly birds). In evolutionary biology we call this convergent evolution. As you might expect, there are also many cases where the eyespots of two species will be similar because they both came from a common ancestor that had eyespots.
As an evolutionary biologist, I am interested in deciphering how many times has a given trait evolved, as well as why species evolve the traits that they do. These questions are tightly linked because the more times a trait arises the easier it is to detect a pattern that provides some insight into the selection pressure that influenced the evolution of that trait. To examine these patterns we use phylogenetic comparative analyses. First, we quantify a character of interest (e.g., presence/absence of eyespots) for several modern-day species, then for the same set of species we collect additional data, such as another trait or habitat feature, that we think might push a species to evolve the character of interest. Next, it is critical when looking for evolutionary patterns that you control for the relatedness among the species you are examining. Failing to account for species’ relatedness invariably results in overestimating the strength of the true correlation between the traits. Again, this is because we need to distinguish between two scenarios: 1) where two species share similar traits because they evolved from a common ancestor that possessed both traits, and 2) where two species share similar traits because evolution acted in a similar way on both species (i.e., convergent evolution). In statistical terms, we need to correct for the lack of independence among related species (i.e., species that share a common ancestor are not statistically independent data points). A variety of techniques are now available to do this, all striving to account for the lack of independence among related species. Essentially, they all boil down to mapping our species of interest, and their associated traits, onto a phylogenetic tree. If many of the species being compared possess the same set of traits because they share a common ancestor (Scenario1) there is less evidence for a relationship between those traits, whereas if many distantly related species share the same set of traits (Scenario2) there is strong evidence for an evolutionary relationship between them.
There is now evidence which indicates that caterpillars are more likely evolve aposematism (i.e., bright colours and associated unpalatablity) when they live and feed in areas where they are highly visible to predators, and not simply because they feed on plants which contain toxins (Prudic et al. 2007 PNAS). But how did the authors figure this out? They started by classifying the colour patterns of existing Papilio genus caterpillars* (i.e., bird-dropping, green, aposematic):
"Green" Type - Papilio canadensis
Next, they tested to see whether the aposematic trait correlated more strongly with the toxins of the corresponding food plant, or the leaf "archetecture" (i.e., broad vs. narrow leaves) of that food plant:
Narrow-leaved - Daucus carota
Of the 59 Papilio species examined, 15 species fed on narrow-leaved plants, 12 were aposematic, and 11 were both aposematic and narrow-leaf feeders. Remember: each species does not represent an independent data point because some share the same set of traits because they share a common ancestor. The authors present genetic evidence which maps their 59 Papilio species onto a phylogeny. It indicates that 7/12 aposematic species examined descended from a common ancestral species that was both aposematic and fed on narrow-leaved plants. In other words, these 7 species possess the same traits (warning colours and narrow-leaved plant feeding) because of an evolutionary event that occurred in their ancestor and not because they were subject to the same selection pressure and independatly evolved the traits. The weight of evidence for the overall pattern is adjusted to account for this. Yet, their analysis also shows that aposematic larvae have evolved independently about 6 times (within the 59 Papilio species examined). The remaining 5 aposematic species were not closely related to each other and 4 of them feed on narrow leaved plants. In contrast, there was no clear relationship between toxins of the food plant and the possession of warning colours. The authors thus conclude that possession of warning colours was more closely correlated to leaf architecture than plant toxin. This is an interesting result and suggests that prey which can't hide effectively from their predators are more likely to evolve toxins and showy colours to advertise their unpalatability.
I am currently using a similar approach to address the evolution of eyespots in various groups of Lepidopteran caterpillars. I hope to have this done for the Evolution2012 conference here in Ottawa, a joint conference between the following societies: American Society of Naturalists (ASN), the Canadian Societyfor Ecology and Evolution (CSEE) the European Society for EvolutionaryBiology (ESEB), the Society for the Study of Evolution (SSE), and the Society of Systematic Biologists (SSB).
*Note Nearly all Papilio caterpillars share a common "bird-dropping" colour pattern until the ~4th instar. Papilio caterpillars with eyespots were classified as "green" in this study.
- Black Swallowtail (Papilio polyxenes) caterpillar
- Field and Lab Notes (more pics of early instar Papilio canadensis caterpillars)