The Panda’s Thumb (24 page)

Haeckel's original illustration of
Bathybius
. The discoidal structures are coccoliths in the gelatinous mass.

Bathybius
and
Eozoon
, ever since, have been treated by scientists as an embarrassment best forgotten. The conspiracy succeeded admirably, and I would be surprised if one percent of modern biologists ever heard of the two fantasies. Historians, trained in the older (and invalidated) tradition of science as a march to truth mediated by the successive shucking of error, also kept their peace. What can we get from errors except a good laugh or a compendium of moral homilies framed as “don'ts”?

Modern historians of science have more respect for such inspired errors. They made sense in their own time; that they don't in ours is irrelevant. Our century is no standard for all ages; science is always an interaction of prevailing culture, individual eccentricity, and empirical constraint. Hence,
Bathybius
and
Eozoon
have received more attention in the 1970s than in all previous years since their downfall. (In writing this essay, I was guided to original sources and greatly enlightened by articles of C. F. O'Brien on
Eozoon
, and N. A. Rupke and P. F. Rehbock on
Bathybius
. The article by Rehbock is particularly thorough and insightful.)

Science contains few outright fools. Errors usually have their good reasons once we penetrate their context properly and avoid judgment according to our current perception of “truth.” They are usually more enlightening than embarrassing, for they are signs of changing contexts. The best thinkers have the imagination to create organizing visions, and they are sufficiently adventurous (or egotistical) to float them in a complex world that can never answer “yes” in all detail. The study of inspired error should not engender a homily about the sin of pride; it should lead us to a recognition that the capacity for great insight and great error are opposite sides of the same coin—and that the currency of both is brilliance.

Bathybius
was surely an inspired error. It served the larger truth of advancing evolutionary theory. It provided a captivating vision of primordial life, extended throughout time and space. As Rehbock argues, it played a plethora of roles as, simultaneously, lowliest form of protozoology, elemental unit of cytology, evolutionary precursor of all organisms, first organic form in the fossil record, major constituent of modern marine sediments (in its coccoliths), and source of food for higher life in the nutritionally impoverished deep oceans. When
Bathybius
faded away, the problems that it had defined did not disappear.
Bathybius
inspired a great amount of fruitful scientific work and served as a focus for defining important problems still very much with us.

Orthodoxy can be as stubborn in science as in religion. I do not know how to shake it except by vigorous imagination that inspires unconventional work and contains within itself an elevated potential for inspired error. As the great Italian economist Vilfredo Pareto wrote: “Give me a fruitful error any time, full of seeds, bursting with its own corrections. You can keep your sterile truth for yourself.” Not to mention a man named Thomas Henry Huxley who, when not in the throes of grief or the wars of parson hunting, argued that “irrationally held truths may be more harmful than reasoned errors.”

I SPENT
D
ECEMBER
31, 1979 reading through a stack of New York Sunday papers for the last weekend of the decade. Prominently featured, as always in the doldrums of such artificial transition, were lists of predictions about “ins” and “outs” across the boundary: what will the eighties reject that the seventies treasured? what, despised during the seventies, will the eighties rediscover?

This surfeit of contemporary speculation drove my mind back to the last transition between centuries and to a consideration of biological ins and outs at this broader scale. The hottest subject of nineteenth-century biology did suffer a pronounced eclipse in the twentieth. Yet I happen to maintain a strong fondness for it. I also believe that new methods will revive it as a major concern for the remaining decades of our century.

Darwin's revolution led a generation of natural historians to view the reconstruction of life's tree as their most important evolutionary task. As ambitious men embarked upon a bold new course, they did not focus narrowly upon little twiglets (the relation of lions to tigers), or even upon ordinary branches (the link between cockles and mussels); they sought to root the trunk itself and to identify its major limbs: how are plants and animals related? from what source did the vertebrates spring?

In their mistaken view, these naturalists also possessed a method that could extract the answers they sought from the spotty data at their disposal. For, under Haeckel's “biogenetic law”—ontogeny recapitulates phylogeny—an animal climbs its own family tree during its embryological development. The simple observation of embryos should reveal a parade of adult ancestors in proper order. (Nothing is ever quite so uncomplicated, of course. The recapitulationists knew that some embryonic stages represented immediate adaptations, not ancestral reminiscences; they also understood that stages could be mixed up, even inverted, by unequal rates of development among different organs. Yet they believed that such “superficial” modifications could always be recognized and subtracted, leaving the ancestral parade intact.) E.G. Conklin, who later became an opponent of “phylogenizing,” recalled the beguiling appeal of Haeckel's law:

But the turn of the century also heralded the collapse of recapitulation. It died primarily because Mendelian genetics (rediscovered in 1900) rendered its premises untenable. (The “parade of adults” required that evolution proceed only by an addition of new stages to the end of ancestral ontogenies. But if new features are controlled by genes, and these genes must be present from the very moment of conception, then why shouldn't new features be expressed at any stage of embryonic development or later growth?) But its luster had faded long before. The assumption that ancestral reminiscences could always be distinguished from recent embryonic adaptations had not been sustained. Too many stages were missing, too many others discombobulated. The application of Haeckel's law produced endless, unresolvable, fruitless argument, not an unambiguous tree of life. Some tree builders wanted to derive vertebrates from echinoderms, others from annelid worms, still others from horseshoe crabs. E.B. Wilson, apostle of the “exact,” experimental method that would supplant speculative phylogenizing, complained in 1894:

Phylogenizing fell from general favor, but you can't keep an intrinsically exciting subject down. (I speak of high-level phylogenizing—the trunk and limbs. For twigs and small branches, where evidence is more adequate, work has always proceeded apace, with more assurance and less excitement.) We didn't need “Roots” to remind us that genealogy exerts a strange fascination over people. If uncovering the traces of a distant great-grandparent in a small overseas village fills us with satisfaction, then probing further back to an African ape, a reptile, a fish, that still-unknown ancestor of vertebrates, a single-celled forebear, even to the origin of life itself, can be positively awesome. Unfortunately, one might even say perversely, the further back we go, the more fascinated we become and the less we know. In this column, I will discuss one classic issue in phylogenizing as an example of the joys and frustrations of a subject that will not go away: the origin of multicellularity in animals.

Ideally, we might hold out for a simple, empirical resolution of the issue. Might we not hope to find a sequence of fossils so perfectly intermediate between a protist (single-celled ancestor) and a metazoan (multicelled descendant) that all doubt would be erased? We may effectively write off such a hope: the transition occurred in unfossilizable, soft-bodied creatures long before the inception of an adequate fossil record during the Cambrian explosion, some 600 million years ago. The first metazoan fossils do not surpass the most primitive modern metazoans in their similarity to protists. We must turn to living organisms, hoping that some still preserve appropriate marks of ancestry.

There is no mystery to the method of genealogical reconstruction. It is based on the analysis of similarities between postulated relatives. “Similarity,” unfortunately, is no simple concept. It arises for two fundamentally different reasons. The construction of evolutionary trees requires that the two be rigorously separated, for one indicates genealogy while the other simply misleads us. Two organisms may maintain the same feature because both inherited it from a common ancestor. These are
homologous
similarities, and they indicate “propinquity of descent,” to use Darwin's words. Forelimbs of people, porpoises, bats and horses provide the classic example of homology in most textbooks. They look different, and do different things, but are built of the same bones. No engineer, starting from scratch each time, would have built such disparate structures from the same parts. Therefore, the parts existed before the particular set of structures now housing them: they were, in short, inherited from a common ancestor.

Two organisms may also share a feature in common as a result of separate but similar evolutionary change in independent lineages. These are
analogous
similarities; they are the bugbear of genealogists because they confound our naive expectation that things looking alike should be closely related. The wings of birds, bats and butterflies adorn most texts as a standard example of analogy. No common ancestor of any pair had wings.

Our difficulties in identifying the trunks and limbs of life's tree do not record muddled thinking about methods. All major naturalists, from Haeckel on (and even before) stated their procedure correctly: separate homologous from analogous similarity, discard analogies, and build genealogy from homology alone. Haeckel's law was a procedure, unfortunately incorrect, for the recognition of homology. The goal is, and has been, clear enough.

In a broad sense, we know how to identify homology. Analogy has its limits. It may build striking external, functional similarity in two unrelated lineages, but it does not modify thousands of complex and independent parts in the same way. At a certain level of precision, similarities must be homologous. Unfortunately, we rarely have enough information to be confident that this required level has been attained. When we compare primitive metazoans with different protists as potential relatives, we often work with only a few features held in common for any contrast—too few to be sure about homology. Moreover, small genetic changes often have profound effects upon external, adult form. Therefore, a similarity that looks too uncanny and complex to arise more than once may actually record a simple and repeatable change. Most importantly, we aren't even comparing the right organisms, but only pale reflections of them. The transition from protist to metazoan occurred more than 600 million years ago. All true ancestors and original descendants disappeared eons ago. We can only hope that their essential, identifying features have been retained in some modern forms. Yet, if retained, they have surely been modified and overlain with a plethora of specialized adaptations. How can we separate original structure from later modification from new adaptation? No one has ever found an unfailing guide.

Only two scenarios have been favored for the origin of metazoans from protists: in the first (amalgamation) a group of protistan cells came together, began to live as a colony, evolved a division of labor and function among cells and regions, and finally formed an integrated structure; in the second (division), cellular partitions formed within a single protistan cell. (A third potential scenario, repeated failure of daughter cells to separate following cell division, has few takers these days.)

At the very outset of our inquiry, we come up against the problem of homology. What about multicellularity itself? Did it arise only once? Have we explained its occurrence in all animals once we decide how it arose in the most primitive? Or did it evolve several times? In other words, is the multicellularity of various animal lineages homologous or analogous?

The metazoan group usually regarded as most primitive, the sponges, clearly arose by the first scenario of amalgamation. In fact, modern sponges are little more than loosely knit federations of flagellated protists. In some species, cells can even be disaggregated by passing the sponge through a fine silk cloth. The cells then move independently, reaggregate into small clumps, differentiate and regenerate an entire new sponge in its original form. If all animals arose from sponges, then multicellularity is homologous throughout our kingdom, and it arose by amalgamation.

But most biologists regard sponges as an evolutionary dead end without subsequent descendants. Multicellularity is, after all, a prime candidate for frequent, independent evolution. It displays the two primary features of analogous similarity: it is reasonably simple to accomplish, and it is both highly adaptive and the only potential path to the benefits it confers. Single cells, ostrich eggs notwithstanding, cannot become very large. The earth's physical environment contains scores of habitats available only to creatures beyond the size limit of a single cell. (Consider only the stability that arises from being large enough to enter a realm where gravity overshadows the forces that act upon surfaces. Since the surface/volume ratio declines with growth, increasing size is the surest path to this realm.)

Not only has multicellularity evolved separately in the three great higher kingdoms of life (plants, animals, and fungi), but it probably arose several times in each kingdom. Most biologists agree that all origins within plants and fungi occurred by amalgamation—these organisms are the descendants of protistan colonies. Sponges also arose by amalgamation. May we then close the issue and state that multicellularity, although analogous both across and within kingdoms, evolved in the same basic way each time? Modern protists include colonial forms that display both regular arrangement of cells and incipient differentiation. Remem ber the
Volvox
colonies of high school biology labs? (Actually, I must confess that I don't. I attended a public high school in New York just before Sputnik went up. We had no lab at all, though it arrived in a flash just as I left.) Some volvoxes form colonies with a definite number of cells arranged in a regular manner. The cells may differ in size, and reproductive function may be confined to those at one end. Is it such a big step to a sponge?

Only among animals may we make a good case for another scenario. Did some animals, ourselves included, arise by division? This question cannot be answered until we resolve one of the oldest riddles in zoology: the status of the phylum Cnidaria (corals and their allies, but also including the beautiful, translucent Ctenophora, or comb-jellies). Almost everyone agrees that the Cnidaria arose by amalgamation. The dilemma resides in their relationship with other animal phyla. Almost all possible schemes have their supporters: cnidarians as descendants of sponges and ancestors of nothing else; cnidarians as a separate branch of the animal kingdom without descendants; cnidarians as the ancestors of all “higher” animal phyla (the classical view of the nineteenth century); cnidarians as degenerate descendants of a higher phylum. If either of the last two schemes can ever be established, then our issue is settled—all animals arose by amalgamation, probably twice (sponges and everything else). But if the “higher” animal phyla are not closely related to cnidarians, if they represent a third, separate evolution of multicellularity in the animal kingdom, then the scenario of division must be seriously considered.

Supporters of a separate origin for the higher animals generally cite the Platyhelminthes (flatworms) as a potentially ancestral stock. Earl Hanson, a biologist at Wesleyan University, has been a leading crusader, both for a platyhelminth origin of higher animals and for the scenario of division. If his iconoclastic view prevails, then the higher animals, including humans of course, are probably the only multicellular products of division rather than amalgamation.

Hanson has pursued his case by studying the similarities between a group of protists known as ciliates (including the familiar
Paramecium
), and the “simplest” of flatworms, the Acoela (named for their failure to develop a body cavity). Many ciliates maintain large numbers of nuclei within their single cell. If cellular partitions arose between the nuclei, would the resulting creature be enough like an acoelous flatworm to justify a claim for homology?