As part of my interest in symbiology I have recently explored the controversial work of Don Williamson on the origins of larvae. Williamson has promoted the idea of radical hybridization being involved in the origins of caterpillars into butterflies - and many other larval forms. For example here is his paper, Caterpillars evolved from onychophorans by hybridogenesis. Basically, Williamson claims that ancestors of modern butterflies may have had their eggs fertilized with sperm from velvet worms. Williamson's work has been widely ridiculed and castigated.
Like many students of symbiosis I am attracted to the possibility of biological metamorphosis arising as a result of fusion between widely separated forms. However, I think that there are more possible mechanisms than radical inter-species hybridization.
I have long thought that another possibility for the evolution of biological metamorphosis involves extended symbiosis. This idea shares the idea that larvae and adults started out as individuals members of separate species - but doesn't depend on the viability of radical hybrids. In an extended close symbiosis, parties can transfer genes gradually - via viruses or sperm-mediated gene transfer. They can also assimilate their partner's traits gradually through learning and ordinary natural selection. Radical hybrids are not needed in this kind of scenario - instead evolutionary assimilation can be gradual.
This symbiosis-based theory seems like a clear possibility to me. It holds that at one stage a wasp-like creature planted eggs in a caterpillar-like creature. These parties developed a close relationship and coevolved until one party assimilated the other. Their mutual descendants are caterpillars into butterflies. It is fairly well known that symbiosis promotes horizontal gene transfer. Mutualism, or at least mutual dependence, probably increases its likelihood.
The symbiosis hypothesis would be boosted by discoveries of wasps that have evolved mutualisms with their egg incubators. Wasps are commonly parasites and their incubators are destroyed my multiple wasps during wasp reproduction. However if a relationship develops in which one wasp hatches from one host, the situation starts to look a bit more like the caterpillar into butterfly metamorphosis scenario. Such cases are in fact known - for example, see here for an example involving a single wasp egg per incubator. Exactly how parasitism might turn into mutualism in this case is not obvious - but there are plenty of other cases where parasites have evolved into benign partners and then into obligate mutualists.
Here's another example of one wasp-per host:
Like Williamson's idea, this theory would be boosted by genetic evidence which supported gene transfer between two species. However, since it is not obvious what the ancestral species were, such evidence may remain elusive. This theory doesn't depend on such evidence existing - maybe no gene transfer was involved and one partner assimilated the other one via learning and natural selection. That makes the theory harder to refute - which is not normally considered a virtue among scientists. However, I think we need an alternative to radical hybridization that preserves the idea of separate origins - which itself is strongly suggested by the phenomenon of metamorphosis, according to multiple lines of evidence.
- Syvanen, Michael and Ducore, Jonathan (2010) Whole genome comparisons reveals a possible chimeric origin for a major metazoan assemblage
- Syvanen, Michael (2012) Evolutionary Implications of Horizontal Gene Transfer
- Syvanen, Michael; Ericksen, Bryan; Linz, Simone and Ducore, Jonathan (2014) Quartet Partitioning Reveals Hybrid Origins of the Vertebrate