Or perhaps one should ask, how did butterflies start to become caterpillars?

Some life forms go through stages that bear no resemblance to each other. The caterpillar, for example, bears little resemblance to the butterfly that it becomes. How did that process evolve?

Recently, two scientists suggested that perhaps the genomes of two species could somehow fuse. The Origins of Larvae, a recent article by Donald I. Williamson and Sonya E. Vickers in American Scientist (November-December 2007), advances the possibility that "mismatches between the forms of adult animals and their larvae may reflect fused genomes, expressed in sequence in complex life histories." Fused genomes is a fairly radical idea, as the abstract admits:

Larvae, the immature forms of many animals, are distinct from adult forms by definition. In many life histories—caterpillars and the trochophore larvae of clams and sea snails are examples—larvae and adults bear no resemblance to each other. Biologist Williamson has proposed that larvae are juvenile forms acquired through hybridization—the fusing of two genomes, one of which is now expressed early in an animal's life, the other late. This hypothesis, which goes against traditional thinking that branches on the evolutionary tree cannot fuse to form chimeric species, is one of several possible solutions to open questions about the evolution of larvae. Although an experiment did not yield convincing DNA evidence, the hypothesis is consistent with certain patterns seen in the distribution of genes across species. Along with other evidence of cross-species hybridization, it implies a pattern of evolution that looks more like a network than like Darwin's tree of life.

In other words, the question is so far from being resolved in conventional thinking about evolution that these researchers contemplate replacing Darwin's "tree of life" - the governing paradigm - with a network. And they suggest this even though an experiment to detect fused genomes "did not yield convincing DNA evidence."

To understand why some scientists suggest radical theories like fused genomes, let's look at some other available explanations. Evowiki, for example, offers several Darwinian explanations:

- Many insects undergo fewer stages of metamorphosis than butterflies do. True enough, but the radical transformations of butterfly metamorphosis are not really accounted for by observing that many insects do not undergo them. In fact, it makes the question even more of a puzzle.

- Also, caterpillars are said to be "precocious embryos" that hatched before they assumed the adult form. But how did the original "precocious embryos" survive, when early hatching usually means early death. Also, was there ever a point at which butterflies emerged fully formed from the egg?

- Another attempted explanation is that a larval stage evolved because the species as a whole benefits from the fact that the adults and larva are not in competition. For example, caterpillars and butterflies are not in competition with each other for food because, generally, caterpillars eat leaves and butterflies drink nectar. Thus, the existence of a huge number of caterpillars - who do not reproduce while in that stage - does not threaten the food supply of butterflies, who do reproduce. However, explaining why a given change may be a benefit does not tell us how it happened. It helps us understand why the species did not go extinct as a result of the change.

Part of the problem is that we do not have a fossil record for the details of how butterflies got started. This area is so far from being understood that it prompts radical theories.

This article helps us understand the scope of the adaptations to be explained, and it's only about how butterfly wing scales can be an efficient three-layered mirror.