PSA from Dr. Pangloss

For example epigenetics ...

Evolution -- the theory so malleable that every failure is really just a success in disguise. 

Paging Dr. Pangloss

20 years ago...

Neo-Darwinism has failed as an evolutionary theory that can explain the origin of species, understood as organisms of distinctive form and behaviour. In other words, it is not an adequate theory of evolution. What it does provide is a partial theory of adaptation, or microevolution (small-scale adaptive changes in organisms). It is partial in two senses. First, Neo-Darwinism assumes random genetic variation followed by selection, whereas there is now evidence for a role of directed mutation in adaptive response. That is, genes can evidently respond to environmental circumstances by non-random, adaptive mutation. And second, many of the adaptive "explanations" advanced for biological characteristics simply cannot be taken as serious science. In 1979, Steven Jay Gould and Richard Lewontin published a classic paper that ridiculed much of the adaptationist literature as constituting a "panglossian paradigm", Just So Stories of such dubious scientific value that they discredit the subject. For a few years after that, adaptationists watched their p's and q's more closely. However, the salutary influence of that paper has unfortunately diminished to the point where Just So Stories are again proliferating wildly. A recent example is why the hour-glass shape in women is an adaptive trait, determined by genes. Men select women with large hips and breasts because these are indicators of reproductive potential, or at least men think they are. Women who satisfy these criteria but do not have a small waist are simply fat, which, we are told, is not a good indicator of reproductive potential. Hence the selection of the hour-glass shape. You might think I overheard this in a pub, but it is in fact advanced as a serious proposition by Matt Ridley in The Red Queen following the original proposal by Low, Alexander, and Noonan in Ethology and Sociobiology. If this is science then Rudyard Kipling was a great scientist.  -- Brian Goodwin, biologist, "Neo-Darwinism Has Failed as an Evolutionary Theory"

Dauber Ant-Bees and Mutant Ninja Turtles

In a 2013 study, some biologists purported to finally have "a single well-resolved and strongly supported tree, robust to multiple methods of phylogenetic inference. Apoidea (spheciform wasps and bees) and ants are sister groups, a novel finding that contradicts earlier views that ants are closer to ectoparasitoid wasps. [emphases mine]"  As with so many phylogenetic studies seeking to marry modern genomics with paleontology and morphology, the work involves crafting a best fit for contradictory indications from various genes--which means that many genes must be sampled.  As explained in a NY Times article:

[Ward] and his colleagues then sequenced more than 300 genes from one species of ant and 10 species of bees and wasps, each representing a different major lineage. In the end, they produced a database far larger than in previous studies.  . . . The scientists then used computers to search for an evolutionary tree that best accounted for their genetic data [i.e. that exhibited the fewest contradictions and anomalies]. Their research clearly pointed to mud dauber wasps and bees as the closest cousins to ants.

Mud dauber wasps! Well, how do you like that?!  Well, this new tree needs a new story.

The new study suggests that mud dauber wasps may offer clues to the evolution of ant societies. The female mud daubers don’t just lay an egg on the ground and fly away. Instead, they first create a mud cylinder in which to house the egg[.] . . .
Dr. Ward doesn’t think it’s a coincidence that the new study reveals bees as having evolved from mud-dauber-like ancestors as well. Bees have also evolved complex societies, in which workers fly off to find pollen and nectar to feed their hives. It may be that only a certain kind of wasp could give rise to either kind of highly social insect.
Dr. Boomsma agreed with this idea. “Colony life with altruistic helpers could only evolve after nests and parental care had evolved first,” he said.

It seems so obvious in retrospect.  Dr. Boomsma sounds like Dr. Pangloss in Eldridge and Gould's "Spandrels" essay.  It may well be that it takes very particular wasp adaptations for complex behavior to emerge.  It may well be that it takes very particular adaptations in any clade.  But if it was never obvious that mud dauber behavior held the key, there must be a reason why it could seem so obvious now, when there wasn't a deep biological principle leading us specifically to mud dauber wasps.

Turtle and Wasp

Scientists instead looked to presumably more remotely related wasps that were much more social.   But eusociality is a paraphyletic phenomenon in Hymenoptera.  Ants and bees seem to be very closely related genetically, and it seems likely that much of their shared genomic information enables their shared social complexity.  The more social vespids have been the usual suspects. "Ants and bees are surprisingly more genetically related to each other than they are to social wasps such as yellow jackets and paper wasps, a team of University of California, Davis, scientists has discovered [emphasis mine]," as another article put it, and consequently the "fossil ant" Cariridris bipetiolata can now be reinterpreted as an ancestral sphecid wasp with putative mud dauber behaviors.  Nothing evolves as fast as evolutionary history.

"That's not a boyd, that's a toytle! Isn't it, Moytle?"

According to Nick Crawford (as reported by Science Daily) "Turtles have been an enigmatic vertebrate group for a long time and morphological studies placed them as either most closely related to the ancestral reptiles, that diverged early in the reptile evolutionary tree, or as closer to lizards, snakes, and tuataras." The same article summed up the history of the turtles evolving story with

Paleontological and morphological studies place turtles as either evolving from the ancestor of all reptiles [branching off very early] or as evolving from the ancestor of snakes, lizards, and tuataras [branching off after the dino-croco-avian (i.e.archosaurian) ancestors have branched off]. Conflictingly, genetic studies place turtles as evolving from the ancestor of crocodilians and birds. [emphasis mine]

So, just by looking at the bones, and not the DNA, turtles appeared to be something very different from what they were inside. Everything up until at least the 1990s pointed to a very, very different lineage that DNA evidence would suggest.  In fact, originally they weren't even thought to be diapsids.  According to the Wikipedia summary, "it was later suggested the anapsid-like turtle skull may be due to reversion rather than to anapsid descent. More recent morphological phylogenetic studies with this in mind placed turtles firmly within diapsids, slightly closer to Squamata [snakes & lizards] than to Archosauria [dino-croco-avians]."

With the advent cladistics and molecular studies, while turtles look primitive, in terms of their genetic information they were believed much more dinosaurian than they look.  Still, some molecular studies were combined with morphological studies to argue for anapsid groupings in "Molecules, morphology, and the monophyly of diapsid reptiles" (2001) by M.S.Y. Lee.

But on 7/26/2011,  Science Daily had announced that microRNAs confirmed the morphological assessment in favor of snakes and lizards:

For decades, paleontologists and molecular biologists have disagreed about whether turtles are more closely related to birds and crocodiles or to lizards. Now scientists have developed a new technique using microRNAs for classifying animals, and the secret is out. Turtles are closer kin to lizards than crocodiles. . . . [T]wo scientists from the Mount Desert Island Biological Laboratory [and other collaborators]. . . have developed a new technique using microRNAs for classifying animals, and the secret is out. Turtles are closer kin to lizards than crocodiles.

So, microRNAs, being stable indicators of group evolution according to the researchers, show that previous studies need to be evaluated. ". . . [W]e had the evidence we needed to say that turtles are a sister group to lizards and not crocodiles."  MicroRNA says that turtles speak Parseltongue.

Not much more than a year before, "a morphological study [in spite of genetics?] conducted by Lyson et al. (2010) recovered them as anapsids most closely related to Eunotosaurus"!

A study reported by Science Daily reported on 5/23/2012 "challenge[d] previous anatomical and paleontological assessments," but judging by the 2011 study, it challenged some previous genetic assessments as well.  So why does the microRNA differ so pointedly from studies based upon a genome-wide DNA inference based on "1145 ultraconserved elements (UCEs) and their variable flanking DNA"?  These later studies, of course, use Bayesian methods (or similar likelihood methods) to determine, based on the selected data, what phylogenetic tree cause the fewest problems.   

Science Daily announced on 11/24/14 that "[s]cientists [at Boston University anyway] place turtles in the newly named group 'Archelosauria' with their closest relatives: birds, crocodiles, and dinosaurs."

Research results, appearing in Molecular Phylogenetics and Evolution, describe how a new genetic sequencing technique called Ultra Conserved Elements (UCE) reveal turtles' closest relatives across the animal kingdom. The new genetic tree uses an enormous amount of data to refute the notion that turtles are most closely related to lizards and snakes.  

What a persistent notion.  But then, we were sure for a long time based on paleontology and morphology, what the tree looked like, even though we really didn't know, it turned out... several times.  These evolutionary ninjas repeatedly have eluded capture until recently (we presume).

Parham notes that studying turtle fossils -- particularly the physical features of their bones -- hasn't always painted an accurate evolutionary picture of turtle relationships across continents and through time. "The turtle tree of life based on fossil turtle anatomy didn't match up with the timing of their appearance in the fossil record, as well as their geography," Parham says. "But the tree of life generated at the Academy's CCG is consistent with time and space patterns we've gathered from the fossil record. These new testing techniques help reconcile the information from DNA and fossils, making us confident that we've found the right tree." [emphasis mine]

So now that there is a lot of data to choose from, it is possible to choose the data and the inference weights to come up with a tree that more closely matches the fossil record.  What a relief!  Like its archosaur cousins the crocodilians, the chelonian seems to be a somewhat devolved critter.  With its devolved anapsid skull and its (convergently!) ankylosaurian body form, it's practically the platypus of the diapsids!