08 September 2007

Not-quite-endless Forms Most Beautiful, or Wormholes Through Morphospace

Blogging on Peer-Reviewed ResearchThe jaw-dropping diversity of life begs for explanation, but at the same time it defies description, so much so that it has inspired hyperbole, even hyperbole of the finest variety. (Cf. Annie Dillard's Pilgrim at Tinker Creek.) You might say that the human description of biological diversity has been extravagant, well-nigh endlessly so:

The result of the various, quite unknown, or dimly seen laws of variation is infinitely complex and diversified. It is well worth while carefully to study the several treatises published on some of our old cultivated plants, as on the hyacinth, potato, even the dahlia, &c.; and it is really surprising to note the endless points in structure and constitution in which the varieties and sub-varieties differ slightly from each other. The whole organisation seems to have become plastic, and tends to depart in some small degree from that of the parental type.

-- Darwin, On the Origin of Species, 1st Edition, page 12
And I think everyone's heard the famous last sentence in the Origin:
There is grandeur in this view of life, with its several powers, having been originally breathed* into a few forms or into one; and that, whilst this planet has gone cycling on according to the fixed law of gravity, from so simple a beginning endless forms most beautiful and most wonderful have been, and are being, evolved.
-- Darwin, On the Origin of Species, 1st Edition, page 490
(You may have seen it with three extra words, inserted at the * in the 2nd Edition and all subsequent editions. Some other time.)

Now, I know that Darwin didn't really mean 'infinite' when he wrote 'endless,' but it sure does sound like it, and his awe on contemplating the biosphere is certainly understandable. There are, after all, some really weird organisms out there (there are even 4-leaf clovers). Naturalists, of all people, should be excused for blurting out, "Now I've seen everything." The creations we see are indeed wonderfully beautiful, and their forms seem to be endless.

But they're really not endless, or even nearly so. Of the truly endless forms that are possible in living things, a tiny subset has actually come to be (on Earth, at least).

The 'forms' we're discussing here, by the way, are structures, architectures, what biologists call morphology. A particular type of structure might be called a morph, and the complete set of possible structures would be called a morphospace. A morphospace exists for any type of structure: butterfly wing spots, number of limbs in an animal, twisting shape of a snail's shell.

In fact, sometimes it is the lack of diversity in form that is more remarkable. In other words, sometimes the question faced by evolutionary or developmental biologists is this: why, given all the possibilities, do we only see these few forms? Why not more, or for that matter, why not fewer?

In the 8 June 2007 issue of Science, Prusinkiewicz et al. tackle this very problem, focusing on plant development using tools well-suited to understanding the development and evolution of form. Specifically, they analyzed the development of inflorescences. The inflorescence is the branching section of a plant stem upon which flowers form. Inflorescences take their shape through a process by which the growing part of the stem (the meristem) forms branches that either turn into a flower (and thus stop branching) or turn into another meristem (which can keep forming new branches). It's a pretty simple process geometrically, and perhaps you can see that it could generate an effectively infinite number of possible forms. Even if you don't have access to Science, you can look at their diagram in Figure 1, where you'll see six different types of inflorescence forms. The top three forms (selected, it seems, at random from a gigantic set of possibilities) do not occur in nature. The bottom three represent the three types of forms that are actually seen on Earth. Three! Three?! Three.

Six flowering plants, all with a different type of compound inflorescence. Chromolithograph, c. 1850.
farbendruck und verlag von C. C. Meinhold & Söhne, Dresden

Courtesy of Wellcome Images, Creative Commons license.



What's more, there seems to be some kind of developmental or evolutionary barrier between two of the forms (E and F in Figure 1). Groups of related species rarely display both forms; a given group will be all one form or the other. So in other words, examination of inflorescence architectures reveals two curious aspects of this part of the living world:
  • the vast morphospace of inflorescence structure is almost completely unoccupied, with all known forms crowded into three neighborhoods; and
  • forms are distributed among species in a pattern suggesting the presence of significant constraints on their development.
Before setting out to address these related questions, the authors explain how they began their quest:
Previously, distinct developmental models have been postulated for different inflorescence types leading to a fractured view of phenotypic space. From an evolutionary perspective, however, inflorescence types should be related to each other through genetic changes. A developmental model that encompasses different architectural types within a single parameter space is thus needed.
In other words, the authors are claiming that previous work on these forms has failed to take into account the background of common descent. According to evolutionary theory, flowering plants have a common history, and are therefore all related to each other through common ancestors. The distinct forms we see today must have arisen through modification of the forms from which they descended. And that means that the different forms should arise developmentally through a common set of functions and components, what the authors call a common "parameter space."

In my view, this is an extraordinary example of evolutionary thinking that drives a specific experimental analysis. The authors sought an encompassing developmental model precisely because they noted that the reality of common descent necessitates such a model. So if you've heard that evolutionary theory doesn't make testable predictions or is of no use in modern biology, here's one more demonstration of the falsity of those claims. "Design" considerations sure didn't produce the key insight; on the contrary, the denial of common ancestry that is sadly typical in the ID camp would have precluded the authors' approach.

The authors proceed to craft a developmental model that can account for the growth patterns that yield the three forms. It's elegantly simple, and its power is magnified significantly by the authors' demonstration that the processes involved are accounted for by the actions of known genes.

We could stop here, noting that a complete and testable model for inflorescence development arose from the introduction of evolutionary bases for new hypotheses.

But the authors went one step further. First, note that their model can be visualized as an "exploration" of the inflorescence morphospace (Figure 3). The developmental processes that make up the model drive the form of the inflorescence down just a few potential pathways; myriad alternative forms in the morphospace remain untouched by the "exploration." Now, evolution can be viewed as a similar exploration, not of morphospace but of a fitness space, or a "fitness landscape." ('Fitness' being defined here as advantageous adaptation.) In a fitness landscape, the peaks are forms or morphs of high fitness, as shown in Figure 5. By varying the parameters of their model to reflect known scenarios faced in evolution (different habitats, for example), the authors show that certain forms are optimal under certain conditions. In other words, the highest fitness peak under some environmental conditions, for example, might be form A, while under another set of conditions, it would be form B.

Interestingly, their fitness landscape models provide an explanation for why forms E and F in Figure 1 are nearly mutually exclusive: in the most common fitness landscapes, those two forms occupy peaks that are separated by a low-fitness valley. An evolutionary exploration from E to F would require a trip through an area of significantly suboptimal structure/function relationship. (Look at Figure 5D, for example: the two forms are the red peaks on either side of the landscape, each at the end of a "ramp," separated by a low fitness point at the bottom where the two "ramps" begin.) Not surprisingly, that doesn't happen very often.

Why, though, does it happen at all? Well, the last two graphs in Figure 5 show what is revealed when the fitness landscape incorporates a third dimension, either environmental variation or longevity of the plant. Have a look: the red boxes represent the high-fitness regions of what is now a 3-D fitness landscape. The red paths in those two graphs connect all three of the main forms. In other words, certain evolutionary influences can facilitate the kind of change that seems impossibly unlikely when looking at a simpler 2-D fitness landscape. The authors call the linkages pictured in Figure 5 "evolutionary wormholes," presumably since they seem to represent an unseen path that belies an initial conviction that "you can't get there from here."

If you've gotten this far, and you've heard/read a lot of ID claims, then you should be able to see why I think this article is important to consider in the context of ID challenges to evolutionary theory. After all, "you can't get there from here" is a reasonable paraphrase of a lot of ID challenges to evolutionary theory.

From a simple insistence on an evolutionary viewpoint, to a simple but elegant model for the development of plant architectures, through some molecular genetics, to wormholes through fitness landscapes. Wow! Can "design" do that?

Article(s) discussed in this post:

  • Prusinkiewicz, P., Erasmus, Y., Lane, B., Harder, L.D. and Coen, E. (2007) Evolution and development of inflorescence architectures. Science 316:1452-1456.

No comments: