Uniqueness and Generalization in a Historical Science
Evolution is an opportunistic process that builds solely on its present state. There is a large chance component in the environmental and biological variation that exists at any time and place. Evolution has been a one-time history. Each individual organism inherits, and must work with, the products of the events that happened in its unique prior history. Selection may mold organisms in a given direction now but does not—and cannot—aim toward anything in the future. Evolution does not require (and our theory does not even tolerate) any form of external force, as invoked by some religions, nor does it require internal animistic drive, invoked by Jean-Baptiste Lamarck and Henri Bergson (Bergson 1907), that directs organisms to evolve toward a particular future objective. Lamarck is today famously ridiculed for explaining evolution in terms of traits acquired by the striving of organisms toward some objective during their lives being inherited by their offspring.
But if evolution is contingent and without such directedness, can we expect to find generalizations or just local description? Might there be forms of biological necessity? Can we identify properties of life on Earth that should apply to life that might be discovered elsewhere, or are we simply describing history and calling it theory? In fact, general principles can be identified. Molecular and physical constraints and characteristics can be viewed as universal if we define life as varying,
Genetics and the Logic of Evolution, by Kenneth M. Weiss and Anne V. Buchanan. ISBN 0-471-23805-8 Copyright © 2004 John Wiley & Sons, Inc.
self-replicating chemical interactions. Perhaps more specific to life on Earth are the intrinsic commitments of basing life on carbon and oxygen or on RNA and DNA and proteins.
To identify general principles, we make the core assumption of a single terrestrial origin, look comparatively at various forms of life to identify characteristics they share, and interpret them as mediated by shared historical processes. At a fundamental level, we compare gene coding mechanisms or biochemical usage among diverse species. To obtain ideas about higher-level phenomena, we can compare similar traits in multiple species—for example, forelimbs in mammals and birds. In such instances, we seek partial generalizations that apply to specifiable subsets of life. Many higher-order aspects of complex life are widely shared, even if the details vary, and a manageable number of broad generalizations are possible. Even so, there are usually exceptions because of the contingent, chance element involved in the historical phenomenon we call "life."
Is Evolution a Branching Process?
One of the most profound aspects of our notion of evolution is that it is a divergent or branching process. As shown in Chapter 1, the metaphor of a tree has been used as the fundamental image of evolution, at least since Darwin and Wallace. But is it correct?
It is true that mutation, natural selection, and other processes lead to divergence, or the accumulation of differences among different lineages descending from a common ancestor. If we look forward in time, the process produces a branching relationship as descendants of an individual, species, and so forth are isolated from each other, each acquiring unique new variation. Because of its ad hoc and complex nature and because of sequestration of the players, the process is, essentially, irreversible. If this is how life works, over time it would generate ever-proliferating diversity, with hierarchical, cladistic, or nested variation within branches sharing common ancestors.
In practice, we cannot really look forward in time, but must use the patterns seen among current individuals and species (calibrated by our understanding of muta-tional processes and the limited amount of geological and morphological information provided by the fossil record) to look backward in time. By grouping organisms with shared traits and assuming that to be the result of common ancestry, what we see involves the "coalescing" of today's forms into ever-fewer ancestral forms—ulti-mately going back to the common origin of life. This does not generally imply that there were fewer forms around at that time, except in a very general sense in relation to the very earliest times, because many past life forms will have left no descendants living today.
However, and especially if we attempt to infer the nature of the first life, the picture is less nested and treelike than the usual conception of evolutionary relationships. For example, reconstructing this history from DNA sequence data reveals inconsistent phylogenies among genes, suggesting that some systems have been transferred horizontally among long-separated branches. The recipient branch's original system is thus replaced in its descendants by the system transferred in from a different branch (Doolittle 1998; Doolittle and Logsdon 1998; Jain et al. 1999; Koonin et al. 2000; Ochman 2001; Ochman et al. 2000).As a result, today two groups A and B e.g. may appear to share a common ancestry based on studies of one par ticular gene, but groups B and C seem to have a common ancestor relative to another gene. The tree of ancestry is not the same for each gene. The extent of this phenomenon is unclear (Fitz-Gibbon and House 1999).
Essentially, horizontal transfer from one individual to a peer of a different species is a violation of the general darwinian postulate of variation transmitted from parent to offspring. Subsequently the immigrant gene, acquired horizontally, is passed by the recipient to its offspring in the normal way. Despite the occasional transfer of genes or sets of genes, one might nonetheless surmise that all of this reticulated ancestry at least goes back to a single ancestral cell. But even this may not be so.
An alternative to a single tree of all cells is that life evolved from a kind of communal pot of primordial soup, dished initially into rather imperfectly sequestered troughs, of incompletely different cell prototypes, and finally to more completely separated bowls. In the first stage of such a scenario, biochemical reactions took place communally, diffusing rather freely. As membranes or other barriers developed, some local, isolated specialization evolved tolerant to horizontal transfer (Woese 2000; 2002). Over time, local environments became more highly structured and organized and sequestration more important and impermeable—in the form of cells. Only by the latter stage were cell types among the major branches of life essentially isolated permanently from each other.
Woese (Woese 2002) questions when the effective isolation occurred, noting that basic gene replication mechanisms share little homology between archaea and other cell types. One possible example is DNA transcription and protein translation systems (the nature of these things will be described in Chapter 4), but mechanisms even more fundamental than that may also have transferred horizontally (Jain, Rivera et al. 1999).An incoming gene that cannot be used quickly becomes mutated into extinction, and it must be in the germ line of the recipient species to be transmitted (a major barrier to horizontal transfer in multicellular organisms). However, even among single-celled organisms, horizontal transfer more likely involves mechanisms that depend on the host having cell machinery compatible with that of the donor's.
Bacteria Eukarya Archaea
Bacteria Eukarya Archaea
Eventually, according to this scenario, cells became too metabolically and genetically integrated across their diversity of functions to be tolerant to the importation of genetic mechanisms other than things basically self-contained. A "Darwinian threshold" is reached after which little horizontal transfer can occur, and cells and their descendants generate the standard kind of diverging, diversifying "tree" of relationships. There is, however, still debate about how much actual horizontal transfer has occurred, centering on, among other things, the analytic methods used to identify possible transfer products in present-day cells and species.
Eukaryotic cells (those with nuclei) are thought to have evolved by cell fusion (e.g., Hartman and Fedorov 2002) that introduced the subcellular organelles mitochondria and chloroplasts into a cell. This was initially a symbiotic relationship (a mutually beneficial arrangement between otherwise self-standing entities) but eventually the descendant cells and organelles were unable to survive without each other. Horizontal transfer may have been important in the evolution of algae, enabling them to engulf each other (or algal-bacterial transfer) and thereby acquire new function (Archibald et al. 2003). We also know that horizontal transfer at least occasionally still occurs, for example, by the insertion of viral genes into recipient genomes.
If there were complete sequestration, every cell lineage would be an entirely separate species. In fact, a number of normal mechanisms modulate sequestration and connect branches of life. Sexual reproduction within a species keeps the germ lines connected. This kind of transfer is easy and generally complete because donor (e.g., sperm or pollen) and recipient (egg or ovum) are so similar. Individual genes within species also are kept connected by recombination between an individual's incoming paternal and maternal genomes.
Life of course is generally cladistic, especially with regard to complex organisms. As the barrier to horizontal transfer became basically impermeable, species and their diversification in the usual darwinian sense became possible. This shows the importance of sequestration in biology.
All metaphors limp to some extent, but because various mechanisms of horizontal transfer exist, there is no single tree of life. Instead, what we see is reticulated, or interconnected, networks of historical relationships. This is true of deep aspects of species relationships, and also of gene relationships in shallower time.
Directedness in Evolution: Do Organisms Solve Problems?
It is difficult to write about evolution without using some kinds of convenient verbal shorthand, such as referring to existing adaptations as if they evolved for what they do today or to solve a problem confronted by the organism's ancestors. That is, we often tend to equate today's function with the selective forces of the past. Examples might be the ability to think, or fly, or digest cellulose. A cow's ancestors developed the ability to digest grass. However, no ancestral insectivorous mammal faced a field of grass and pondered how to digest it, the way we face a field of grass and ponder how to mow it. Organisms have no known way to develop heritable means to solve problems identified prospectively.
It is easy to think of environments as presenting problems for organisms to solve. But this can mislead us into Lamarckian thinking. The presence of the atmosphere makes flying possible, and flight has evolved many times. But birds' reptilian ancestors did not have to fly, as many contemporary land-bound reptiles demonstrate.
Nor has flight always taken the same form. An opportunity is not the same as a necessity. We can easily imagine opportunities that have not been taken.
Perhaps, genuine lamarckian mechanisms for directly producing heritable change in response to environmental circumstances will be discovered; some possible instances have been offered. There are examples to suggest that "evolvability" may exist, in that some organisms under stress respond by producing mutations, perhaps even in a context-specific set of genes (Caporale 1999; Fontana, 2002; West-Eberdard, 2003). The idea is that at some point organisms (and here we are speaking of single-celled organisms) had regions of DNA that were subject to mutations under, say, nutritional stress, and those mutations by chance led to tolerance of that stress and hence proliferation. However, the specific mutations themselves in this case are random, not directed to the specific need, and involve simple DNA changes rather than complex adaptations. In fact, organisms with a sequestered germ line are less likely to evolve such mechanisms because the mutations generated under stress would have to be inserted in the germ line and not just be in the body itself, yet it is the body that must survive the stress.
Our current understanding is that evolution is not teleological, that is, it does not work with future objectives in mind. We may some day discover lamarckian means of genetic evolution, but until then we have to hold to our view that mutational change is random relative to need. Neither directed, future-anticipating change nor the inheritance of acquired traits provides necessary explanations for the major functional characteristics of organisms. (Life may, however, someday evolve in a teleological way, if we develop genetic engineering methods to produce ends we envision in advance, such as sheep whose milk contains antibiotics useful for treating human disease.) The modern DNA theory of life, often called the Central Dogma, that a specific gene codes for a specific protein but that the gene's structure is not directly affected by how that protein fares in life (see Chapter 4), is the theoretical guarantee of this lack of lamarckian inheritance. However, this has to do in part with how we define heredity, and there are numerous examples of parent-offspring transmission of acquired traits—one being your ability to read this book.
What is a Trait? What Evolves?
It might seem strange to ask what a trait is, if the whole point of understanding evolution is to explain the diversity of traits in organisms. However, there have been and continue to be debates about what exactly it is that we refer to in this context. If traits are selected for or evolve, what are "traits"? Another way this has been put is this: What is the unit of selection?
This sounds simple but is not a trivial question. There is so much diversity in nature. Anything can be a trait. But if we want to relate our discussion to genes and adaptation, we need to know what we are considering relative to genes. Life cycle is a good example. This would seem to be directly related to the notions of darwinian fitness in the face of natural selection, since those who live longest or reproduce first might be declared the evolutionary winner. Does selection work directly on that or on the processes underlying the result? For example, much has been said of the notion that maximum lifespan is a characteristic of a species. This seems sensible, but does it imply there are genes for the timing of death? Does age at death evolve as a trait? Or is it just that causes of disease, that is, problems in cellular physiology, are screened and the net result is a statistical pattern of ages at death? This
seems most likely (e.g., Finch and Kirkwood 2000); indeed, something we have already stressed is that there appears to be a huge component of chance even in life history events. This can easily seem to fly in the face of the universal appearance of adaptation in nature; but does it?
Is Adaptation a Profound or an Illusory Concept?
The lack of foresight in evolution leaves us as scientists (not organisms) with a problem because so many traits appear to have evolved to "solve a problem." Bats certainly fly! And it is not much of a misstatement that they evolved to fly or that they evolved because their ancestors strove to fly or that selection favored flight. Our theory holds that evolution is opportunistic. Selection screens variation that exists (by chance), favoring some functional variation that may or may not relate to flight but is useful at least in some way.
If an aspect of the environment remains relatively constant over long time periods, traits suitable for increasingly effective use of that aspect can be favored. Over time, the trait's evolution can continue in a generally consistent direction because variation that arises is screened by the same factors. This can effectively canalize (channel) evolution in a persistent way (Wagner et al. 1997). Biochemical constraints (see below) can limit what selection can achieve, and can contribute to what the prominent biologist William Bateson (Bateson 1913) referred to a century ago as "positions of organic stability." There is never any foresight involved, but a steady environment can lead to what appears to be directed evolution.
The resulting teleological illusion is what drove Lamarck, the Argument from Design, and numerous other responses to darwinian explanation. However, it can never be stressed enough that, if suitable variation had not arisen under particular circumstances, we would not have observed, for example, flying organisms today. Nothing we know about the mechanisms in biology suggests that flight arose through foresight or internal drive in any animal in the past or suggests that foresight or internal drive was necessary.
We are also somewhat trapped by the anthropic principle referred to in Chapter 1. Every organism we see today is the descendant of four or so billion years of uninterrupted success. Each has inherited genes that history blessed since the primeval soup. It cannot be otherwise. Critics note that, because of this fact, adaptive explanations verge on tautology because one can always invent an adaptive story that leads from past to present, and such explanations are sometimes applied so unconditionally by biologists as to be scientifically not much more meaningful than a collection of Just So stories because they cannot be verified. The issue might be ameliorated if we tempered these adaptive scenarios by keeping in mind that organisms clearly are not as finely tuned to their environment as is often casually assumed.
Similarly, our adaptive scenarios for complex traits might be tempered if we were obliged to specify how it happened. The human brain is considerably larger for our body size than the brains of other primates. We assume this is an adaptation for mental function. However, some humans have much smaller brains than others with no obvious defect (in behavior or, more importantly, reproduction), and genetic variation can seriously affect brain function without affecting brain size. It is easy to "explain" brain evolution by saying that a change of some very small amount in average size per generation (e.g., 1 mm3) would be sufficient over many thousands of generations to increase brain size of the amount observed comparatively and from the fossil record. But how does 1 mm3 of extra brain volume lead to increased reproductive fitness? If brain size did not increase in a gradual way, how did it increase? And why?
These important questions can be asked about most complex traits. For a scientific principle to have much meaning, or to be persuasive in a given circumstance, there should be some constraint on when and how it can be invoked. Adaptive explanations raise fundamental issues about concepts of causation in biology. This can be seen by the fact that, from Darwin to today, many biologists effectively assume that any complex trait is mainly the result of steady, gradual selection. (Rapid change by sudden mutation leading to a new level of complex organization has seemed impossible except for certain special cases, such as segmented, serially homologous systems like hands and feet or legs and wings in insects, in which the number of elements might under restricted circumstances change quickly and in simple ways we now understand at the gene level.)
There is no satisfactorily provable way out of the teleological illusion, but this has not shaken biologists into eschewing the making of adaptive scenarios, mainly because a good enough alternative material explanation for directed change does not exist. Religious creationists scurrilously misrepresent what biologists mean when they say that evolution is due to chance. But ironically, in insisting on adaptive scenarios, biologists share with religious creationists the belief that complex traits cannot arise just "by chance." However, we will suggest below that chance may be a more important factor in adaptive evolution than has been thought.
natural Selection and the Species Question
One of the central concerns in the early stages of systematic professional biology in the late 18th and early 19th century was the "species question," that is, explaining the existence of the diversity of species, each suited to its way of life. Everyone knew that plants and animals were variable, and breeders could modify that variation up to a point. Domesticated species could be bred to change, but when the breeder's attention lapsed they seemed to "revert to type," and breeding never extended to the production of new species. Creationist explanations were weakening, as evidence from fossils, biogeography, and systematic, comparative, anatomic, and taxonomic studies accumulated. These studies showed that life was some type of historic phenomenon, and evidence showing that species did change and that new species arose by diversification from earlier species increased. But how?
Darwin and Wallace provided a general, codified, plausible, and in a sense observable mechanism—natural selection—by which species could in principle arise and change (Patrick Matthew also expressed the same argument clearly in 1831, but it went unnoticed in most subsequent priority credits because it appeared as only a brief comment in an appendix to a paper in a specialized book on naval arboriculture). But this lengthened the prevailing sense of time and made it a critical factor concerning relationships among species. Previously, when evolution referred to development, time was on the embryological scale. Charles Lyell, James Hutton, and other geologists had discovered slow processes by which the Earth's shape is changed, and that was an important factor lending plausibility to the ideas forming in Darwin's mind. His theory required a lot of time, and he was concerned that there might not have been enough for natural selection to mold the wonderful and complex diversity seen on Earth. He was convinced that the biblical estimates of the age of the Earth, roughly 6,000 years, were incorrect, but he thought "We have almost unlimited time ... there must have been ... millions on millions of generations" (C. Darwin, 1858, paper announcing evolution read before Linnean society; available on the public domain and web).
At the time, it was impossible to know just how many millions of generations the age of the Earth might truly have supported, and Darwin struggled with the problem. In fact, he thought that hundreds of millions (perhaps 300) of years would be required, and was highly discomfited by the British Royal Astronomer Lord Kelvin's estimate that the Earth was only about a tenth that old.
We now know that the Earth is much, much older, and that life has been here for several billion years (perhaps even 4 billion). But is that "long enough," not just for the evolution of cells or butterfly wings but for the evolution of all traits in all species, from leaves to language, without exception? Does 4 billion years make complex evolution, or adaptive explanations more or less plausible than some younger age? This is really a moot point, which is why Darwin could persist in his views despite unclear and sometimes quite contrary arguments about the age of the Earth. There is no real way to know how long is long enough for selection to have done its job. So long as we explain adaptation conditional on the assumption that life has evolved by natural selection on Earth, it must be old enough—it was old enough! Debates about this today are usually waged over contending adaptive scenarios. But so long as we accept the theory, there is no issue of the adequacy of time; instead, it is our job as biologists to use our theory to understand the details of how a given being evolved during its respective historic interval, which we document, for example, by molecular "clocks," calibrating time by the number of mutations that occur in DNA sequences compared among several species, and from fossil dates and biogeography. As we will see in Chapter 3, general mathematical and statistical theories have been developed for the rate and rapidity of selection in populations under various specified characteristics, and, while oversimplified, this yields an understanding of the way the process works in principle.
However, we will see in the next section that in some sense the ancient age of the Earth may actually make our reliance on natural selection less necessary than we have generally thought, which may change the kinds of reconstructions we should make or, at least, the need to invoke selection as much or as determinatively as we do.
Was this article helpful?