Pain is a great teacher, I tell audiences. Take it from me. If PZ Myers hadn’t designated April 7 "Paul Nelson Day" a few years ago, to lampoon me for my failure to explain the concept of "ontogenetic depth," I would never have learned just how intractable the problem of animal macroevolution would turn out to be.
"Oh, come on, Paul — you’re exaggerating," says the skeptical reader.
Not really. Before PZ’s critique, ontogenetic depth (OD) seemed pretty obvious to me. The metric could be calculated as a straightforward product in any animal species, by multiplying the number of adult cell types by the number of cell divisions, from fertilization and first cleavage onward, yielding a good estimate for comparing developmental complexity among the animals. Smaller animals with fewer cell types should exhibit a lesser degree of OD than larger animals with more cell types. Easy, right?
Easy, that is, until one actually tries to calculate the value. And that’s where PZ’s original critique sent me off on a long path that continues today. I see more clearly than ever why the origin of developmental pathways requires a cause with foresight.
In presentations and writings from the mid 1990s until PZ’s original OD critique, I had tended to say that the origin of animal body plans was mainly a question of (1) increasing cell number, and (2) increasing cellular differentiation. While both of those points are still true, they don’t come close to touching the main problem. Whatever caused animal body plans to arise had to know where it (namely, the cause) was going. And the first step on that road is the hardest to take.
Given that the origin of animal body plans and the origin of animal development are intimately connected, this means that the problem of animal macroevolution will not be solved using the current limited toolkit of evolutionary theory. Foresight is a teleological, or design-based concept, and thus verboten (for philosophical, not evidential, reasons) at the moment in evolutionary biology. Concepts that are off-limits need a change in philosophy before they can be reintroduced into a discipline.
But you know what? You can see this for yourself. Try the following exercise.
On the Origin of the First Cleavage Stages in the C. elegans Worm
Caenorhabditis elegans is a model system about which biology has learned a great deal over the past forty years. Compared to mammals, or even fruit flies, C. elegans is a relatively simple animal, with 1,031 cells in its adult male phenotype (959 in the hermaphrodite), apportioned into a variety of specialized cell types and tissues within the tapering nematode body plan. C. elegans was the first animal to have its entire genome mapped, and remarkably, the developmental lineage of every cell in the adult worm has been painstakingly tracked (work that won the Nobel Prize for John Sulston). From the perspective of detailed biological knowledge, it’s hard not to fall in love with these little worms.
If one looks up "origin and evolution of nematodes," however, one will find papers on the phylogeny of the clade — but all such papers presuppose the existence of the nematode body plan. (Some people find "body plan" a bothersome and unhelpful concept, laden with typological baggage. But as you’ll see, the problem we’ll examine would exist whether we placed C. elegans in Phylum Nematoda or not.) What one won’t find are any papers showing how the nematode body plan itself came to be, from eukaryotic unicellular or colonial ancestors. And, as you’ll see from the exercise below, there’s a good reason for that.
Since "body plan" may seem too abstract, let’s stick with well-understood C. elegans. Following fertilization, the first event in C. elegans development is the establishment by cell cleavage of the major founder lineages (see Figure 1). This cellular branching pattern, characteristic of C. elegans, is remarkable in many respects, but we should focus on just a couple of aspects. As a shorthand, let’s call this character CEICP (for "C. elegans initial cleavage pattern").
Fig. 1. The initial cell cleavages following fertilization in C. elegans. AB and P1 are the primary daughter founder cells, giving rise to the AB, MS, and C lineages (containing mixtures of ectodermal and mesodermal cells), D (muscles), E (intestine), and P4 (germ cells). Abbreviated CEICP in text.
First, let’s consider the evolutionary framework for the puzzle. Like all animals, C. elegans must have descended from unicellular eukaryotic ancestors, perhaps via — well, there is the mystery.
Figure 2 shows the road on which the evolutionary processes at the origin of C. elegans must travel, where the distance marker is increasing cell number. On the left, the starting point, is the unicellular eukaryotic state. On the right, our destination, lie the approximately 1,000 cells of the adult C. elegans.
Fig 2. The distance, measured in terms of cell number, between an ancestral single-celled eukaryote and adult C. elegans. Along this evolutionary branch, cell number must increase.
Okay, the question for the exercise: Where on this interval did CEICP first evolve?
Saint Denis Carrying His Head, and Evolving Initial Cleavage Patterns De Novo
"La distance n’y fait rien; il n’y a que la premiere pas qui co�te," observed Marie Anne de Vichy-Chamrond, marquise du Deffand: "The distance is nothing; it’s only the first step that counts." She was a wise woman. Writing to d’Alembert in 1763 about the miracle of St. Denis — who, according to legend, after his execution by decapitation at Montmarte carried his head for several miles through Paris, preaching a sermon as he went — the Marquise observes that the entirety of the feat resides in the first step. The rest is routine.
So let’s try to take the first step in the origin of CEICP, by placing the character at its correct location (i.e., the point of first appearance) on the phylogenetic road from single-celled eukaryote to adult worm.
Obviously CEICP cannot evolve very late (see Figure 3) — anywhere in the neighborhood of 1,000 cells — because CEICP is necessary to specify the terminal fates of those 1,000 cells. No CEICP, no worms.
Fig. 3. Origin of CEICP when total cell number is closer to 1,000 than 1.
So let’s move CEICP to the other end of the interval, much closer to fewer cell numbers. (See Figure 4.) How about here?
Fig. 4. Origin of CEICP when total cell number is closer to 1 than 1,000.
But now we face two different, but related, problems:
- The exact features of CEICP, whose origin we want to explain — namely, the precise decision-tree logic by which the zygote is subdivided into founder lineages with specific fates — disappear entirely. Those features disappear because the normal functional role of CEICP is end-directed, aiming towards the adult worm, and the adult worm doesn’t exist yet. Its evolution lies in the distant future. The processes of evolution, whether selection, drift, or some other mechanism, have no foresight.
- As cell numbers drop towards the single-celled state, it is unclear that functional cell types and tissues will continue to exist. Again the difficulty is descriptive. "Simpler" animals, such as Trichoplax, with only a handful of cell types, actually possess many more cells than C. elegans. The paucity of detailed anatomical or developmental descriptions of metazoan ancestors, for C. elegans or any other existing animal species — beyond cartoon or nondescript "schmoo" drawings, at any rate — testifies to the challenge of describing genuinely functional organisms where the total cell number of the adult has significantly decreased. Take away enough cells, and again, there’s nothing left to explain, at least that we can actually describe.
In short: no worms, no functional need for CEICP.
So whatever CEICP was doing when it first came to be had nothing whatsoever to do with its current role.
This may not seem troublesome to evolutionary thinkers accustomed to explaining by using concepts such as "exaptation." Yet we still need to describe CEICP, because the character needs to evolve somehow in the phylogeny of C. elegans, at some point in the interval, and now description is impossible. The very features of greatest interest to us have been scrubbed away. There is nothing left to explain.
If Saint Denis carried his head through Paris, that would have been a bona fide miracle — but really, only his first step mattered. If CEICP evolved via an undirected evolutionary process, that developmental character must have come into existence by violating what is thought to be the case for natural selection, or drift, or any other realistic evolutionary mechanism. It’s the first step that counts in explaining the origin of any developmental pathway, because everything downstream relies on the starting point.
And the first step remains unexplained.
This Problem Lives Everywhere in Explaining the Macroevolution of Animals
Ontogenetic depth was my first attempt to grasp how and why animal developmental pathways showed varying degrees of complexity, and to measure those differences. The attempt failed (OD-ed, if you will) because the phenomena in question go well beyond the crude metrics of cell type and number of divisions from fertilization.
Thanks to PZ Myers’s annual prodding, however, I have thought much more deeply about the problem of evolving animal development than I ever expected to do. And, in searching the literature (indeed, in conversations with PZ himself, at Society for Developmental Biology meetings), I’ve found that current attempts to solve the problem fall hopelessly short of the mark. I’ve focused here on C. elegans, because the species is (relatively) well understood, but the same general difficulty applies to the origin of any animal group. Go into the literature yourself, and you’ll see what I mean.
Current evolutionary theory falls short because it excludes a priori notions such as foresight — not for any evidential reason, but because foresight requires mind, and mind is philosophically unacceptable within the prevailing materialist outlook.