The Twin Nested Hierarchy

Post of the Month: April 2000

by Ken Cox

Subject:    Re: Can you prove evolution scientifically?
Newsgroups: talk.origins
Date:       April 13, 2000
Message-ID: 38F6071E.8ABDF9C3@research.bell-labs.com

"Bradley V. Stone" wrote:
> Larry Handlin:
> > We do see the nested hierarchy of species dated both relatively and
> > absolutely--how else do you explain this evidence?
> Again, you say we do see this, but please give me an example or two.
> And not a fox is a pug. Or an eagle is a penguin.

Your examples trouble me somewhat, since they suggest that you don't understand the term "nested hierarchy". This refers to the type of pattern that you get when you consider characteristics of a bunch of different species, and group the species together by how similar they are in the characteristics.

This type of grouping has been done for several hundred years now (in a formal sense; people have been doing it informally, for example recognizing groups like "birds" and "fish", for as long as we can trace). The major effort has used morphology, that is, gross characteristics like whether the species is made of wood or meat, or how many legs it has, or how its teeth are arranged, and so forth.

Using morphology, you find that species form a tree structure. For example, you find that there is a huge group of species that have cells with a nucleus. These are called "eukaryotes". You then find that you can divide these up by whether or not they have mitochondria; and so on.

Eventually you get down to a group of hair-possessing milk-giving amniote-possessing tetrapodal jawed vertebrate notochord-possessing multicellular non-chloroplast mitochondrial eukaryotes, also known as the "mammals". Then, within the mammals you find that you can separate out those that lay eggs, those that have pouches, and those that bear live young; and again you keep going, and eventually you end up with a group of tailless, forward-facing-eyes, grasping-paws, hair-possessing etcs. Then, if you look real close at quite minor characteristics (because the really big characteristics, like having mitochondria and being a vertebrate and giving live birth and so forth, have already been covered), you find that you can separate this group into humans, chimps, bonobos, and gorillas.

Of course along the way you've also defined a lot of other groups that at one point or another don't have a particular set of the characteristics that happened to lead to humans. If you were to start with any of those groups and look at their characteristics you could do the same sort of thing; for example you find that among the non-grasping-paws mammals there is a largish subgroup that has hooves, and within that group there is another grouping that has horns, and within that grouping there are a couple hundred species of antelope that differ only in minor details.

Because you can keep doing this sort of division in each group, you end up with the nested hierarchy structure. That is, at the top level you have a big blob labelled "life". Inside this are some blobs labelled "archaea" and "eubacteria" and "eukaryota". Inside each of those are more blobs. And so forth.

Now, another phrase you might hear is "twin nested hierarchy". The nested hierarchy discussed above is based, as I said, on morphology. But recently we've been able to observe some additional characteristics of species, way down at the biochemical level -- things like the exact sequence of the 16S ribosomal RNA inside the mitochondria inside the cells, or the exact sequence of the cytochrome-C protein. These RNA and protein sequences are shared by a lot of organisms -- or rather, the functions that they carry out are shared, because the sequences can vary quite a lot.

This variation is possible because only a few parts of the RNAs or the proteins are absolutely constrained by what they do. The rest can vary, and indeed in a few cross-species molecular transplantation experiments we find that the variations work just fine -- that is, a wheat cell will work just fine if it has the mouse gene for producing cytochrome-C, and similar stuff.

Now, we can apply the same sort of similarity analysis that we use with bones and teeth and nails to these sequences (actually, it's a bit easier in one way, because there's less debate about the measured characteristics). When this is done, we get another nested hierarchy.

Or rather, we get the same nested hierarchy. That is, when you look at morphology, you find that humans are closer to mice than to fish, and that we're closer to both of them than to wheat. And when you repeat the exercise with the sequences, you get the same set of relationships.

Now, this didn't have to happen -- as noted, the wheat would have worked just fine with the mouse's protein, and vice-versa, and in that case we would have found that humans are closer to wheat than mice. So there are many, many possible trees that could have resulted from the sequence data. But the one that did result is the same as the one that we get using morphology.

There is an easy explanation for this using the theory of evolution. There is not an easy explanation using notions of design -- or rather, all the explanations end up saying that the designer happened to choose, out of all the possible patterns, the one that also makes it look like evolution happened.

-- 
Ken Cox                  kcc@research.bell-labs.com

[Return to the 2000 Posts of the Month]