Motors of Stasis and Change: 
The Regulation of Genetic Stability 

From The Century of the Gene - Evelyn Fox Keller HUP 2000

If the Mendelian revolution marked the turning point of twentieth-century biology, then surely the Darwinian revolution was the great watershed of the nineteenth century. The realm of living organisms could no longer to be fitted into a great "Chain of Being"; it required its own figuration: more of a tree than a chain, and as much a succession of becoming as of beings. The living world became a world in time, and both its occupants and its relational structure were reconfigured as products of its evolutionary history. After the publication of On the Origin of Species in 1859, few could be found among the scientifically literate who still believed in the fixity of species. Moreover, Darwin's evolutionary theory offered his readers a mechanism for the origin and transformation of species-natural selection acting upon individual variation. Yet, for all the power of that theory, a fundamental mystery remained. If change is the essence of life, how are we to account for the remarkable stability with which, in each generation, organisms develop and grow true to the type of their particular species, and with a certainty that endures over the lifetime of that species?

Viewed from the perspective of geological time, species transform and evolve. Yet viewed from the perspective of historical time, they display an unmistakable constancy in form and function. But on this matter-on the "stability of type" (to borrow a phrase from Francis Galton) that is so conspicuously maintained over the course of generations, Darwin's theory was silent. However eloquently and powerfully the theory of evolution by means of natural selection might account for changes in biological form and function occurring over eons and reflected in the geological record, it could not begin to explain the reproducibility of that same form and function over the shorter spans of genealogical time. Nor could it offer any account of the persistence of particular individual features from generation to generation, of the clearly recognizable family resemblances that are passed on from parents to offspring.

Of course, Darwin was not privy to the insights of genetics, nor could he have been. He shared with his contemporaries a belief in "blending heredity"-the view that the characteristics of an offspring are, somehow, a blend of the parents' characteristics-but he had nothing to say about how such distinctive features as the Hapsburg lip might endure without dilution. Nor could he offer any kind of answer to the dilemma that was later to plague Schroedinger: How can we understand the reproduction of individual features, generation after generation, with such fidelity as to lend them a "durability or permanence that borders upon the miraculous?"

The fact is that Darwin's preoccupations were different. Throughout his life, he focused his attention on mechanisms of transformation; the mechanisms required for conservation eluded both his understanding and, for the most part, his interest. And while he acknowledged that "our ignorance of the laws of variation is profound" and devoted considerable attention to the ways in which the variation essential to natural selection might arise, nowhere did he express concern about a corresponding ignorance of the laws of constancy.¹

The task of searching for the laws of constancy-that is, of accounting for intergenerational stability-thus fell to Darwin's heirs. Indeed, the century of the gene begins with this task-or more specifically with efforts to account for the persistence of individual traits through the generations. Of course, just as with any collective endeavor, the science of genetics arose out of multiple needs and a variety of different interests, and these have been well chronicled by many historians. My focus here, in Chapter 1, is on the particular force that the search for constancy of individual traits exerted on the origins of the very concept of the gene. A crucial component of that concept, I argue, enters the history of genetics even before the word gene was coined, and it enters with the supposition that underlying each individual trait is a hereditary unit so stable that its stability can account for the reliability with which such traits are transmitted through the generations. In other words, the problem of trait stability was answered by assuming the existence of an inherently stable, potentially immortal, unit that could be transferred intact through the generations.

In the first part of this chapter, I trace the increasing hold this assumption of the intrinsic stability of hereditary elements came to have on geneticists in the first part of the century, its apparent vindication in the middle of the century, and its gradual dissolution over the last few decades. To be sure, genetic stability remains as remarkable a property as ever, and it is clearly a property of all known organisms. The difficulty arises with the question of how that stability is maintained, and this has proven to be a far more complex matter than we could ever have imagined. Furthermore, we will see that the maintenance of genetic stability turns out to be inextricably bound up with the generation of variability. Thus, in the second part of this chapter, I return to Darwin's concerns, taking up the companion issue of transformation and discussing some of the surprising challenges that new research on mechanisms of conservation pose to the simple neo-Darwinian picture of evolution by the cumulative operation of natural selection on randomly generated small mutations.

Finally, a word about the relation between the stability of "type" (that is, the stability with which organisms, in each generation, develop and grow true to the type of their particular species) and the stability of individual traits. For a long time, it was assumed that genes are as capable of explaining the development of individual traits as they are of explaining the development of whole organisms, and therefore that genetic stability sufficed to account for what I will later on in this book call developmental stability. I use the term to refer to the reliability with which organisms of a particular species undergo the passage from fertilization to maturity, generation after generation, each time reproducing a phenotype that is clearly recognizable as characteristic of that "type." Thus, while genetic stability is a property of all organisms, developmental stability is a term primarily applicable to multicellular organisms that pass through embryonic stages of development-that is, metazoan organisms. The differences between these two kinds of stability may be significant, but discussion of such differences must be deferred until after I have said more about the relation between genes and development. Accordingly, in my fourth and final chapter I return to the particular challenges raised in attempting to account for developmental stability.

EXPLAINING GENETIC STABILITY

August Weismann (1834-1914)-one of the great zoologists of the latter part of the nineteenth century-put the problem succinctly: "When we find in all species of plants and animals a thousand characteristic peculiarities of structure continued unchanged, through long series of generations; when we even see them in many cases unchanged through out whole geological periods; we very naturally ask for the causes of such a striking phenomenon . . . How is it that . . . a single cell can reproduce the tout ensemble of the parent with all the faithfulness of a portrait?"²  In these brief remarks, written in 1885, Weismann defined the challenge for a science of heredity indeed, one might read the entire history of genetics as an attempt to answer the question he posed. But Weismann did more than pose the question: he also proposed something of an answer, and the form of his answer helped set the science of heredity on the particular track it would follow for the next sixty years or more.

Whatever the mechanism by which a single cell reproduces the traits of the parent, Weismann assumed the existence of particulate, self-reproducing elements that "determine" the properties of an organism; appropriately enough, he called these elements determinants. This assumption was hardly unique to Weismann -in fact, Darwin himself had hypothesized the existence of some such elements (his gemmules). The Dutch botanist Hugo de Vries, a near-contemporary of Weismann's (1848-1935), also hypothesized the existence of elementary hereditary units. As he wrote, "Just as physics and chemistry are based on molecules and atoms, even so the biological sciences must penetrate to these units in order to explain by their combinations the phenomena of the living world."³ De Vries called his units pangens, a term he introduced in 1889 in an effort to salvage the best of both Darwin's gemmules and Weissman's determinants.

But Weismann assumed more than the existence of elementary hereditary units. In order to explain the remarkable fidelity with which such traits were reproduced generation after generation, he further hypothesized the sequestration of a full complement of these elements in a substance "of a definite chemical, and above all, molecular composition." He called this substance the "germ-plasm" and argued that a germ-plasm, insulated from the ravages of individual mortality, could be transferred, intact, from one generation to another. Thus he wrote, "I have attempted to explain heredity by supposing that in each ontogeny, a part of the specific germ-plasm contained in the parent egg-cell is not used up in the construction of the offspring, but is reserved unchanged for the formation of the germ-cells of the following generation."' Weismann's theory traveled wide and fast. In his influential textbook published only a few years after Weismann's work had appeared in English, the American zoologist E. B. Wilson wrote, "As far as inheritance is concerned, the body is merely the carrier of the germ-cells, which are held in trust for coming generations."⁵

Experimental biology was still in its infancy at the end of the nineteenth century, and Weismann had no way of knowing what these hereditary elements might be. Nor did de Vries, or any other student of heredity at that time. This was a period of grand speculations, and Weismann's were among the grandest. As he explained his philosophy, "Biology is not obliged to wait until Physics and Chemistry are completely finished; nor have we to wait for the investigation of the phenomena of heredity until the physiology of the cell is complete . . . Science is impossible without hypotheses and theories; they are the plummets with which we test the depth of the ocean of unknown phenomena and thus determine the future course to be pursued on our voyage of discovery."⁶ Given how little they had to go on in the way of concrete evidence, it comes as small surprise to find how much (or how sharply) these early thinkers about heredity differed from one another both in their characterization of hereditary elements and in their conjectures about how these elements could impress their various characteristics on the formation of particular cells and tissues. What is more surprising is how much they shared. Underlying all their differences were two enduring articles of faith.

The first of these was that, just as atoms and molecules provided the fundamental units of explanation in physics and chemistry, so too would particulate hereditary elements serve as the fundamental units of biological explanation. These units might themselves be some kind of atom or molecule, or they might be made up of molecules, but the important point was that they were elemental, the primitive units with which the study of heredity must begin.

The second article of faith was closely related, and it held that responsibility for intergenerational stability inhered in the fixity of these material elements, taken either as individual units or in their collective composition. For Weismann, the burden of stability lay in the sequestration of a certain substance "of a definite . . . molecular composition" in a protected lineage of germ cells, where they would be held inviolate for future generations. For de Vries, it lay in the sequestration of the individual particles in the nucleus of each and every cell, with one particle representing one hereditary characteristic. But once sequestered, whether in the germ-plasm or in the nucleus, the fixity of the elements themselves was simply taken for granted, accepted as part of their definition.

The rediscovery of Mendel's rules of inheritance in 1900 marked the beginning of an end to the era of grand speculation in the study of heredity. Indeed, Johannsen's aim in coining the term gene in 1909 was to mark a break with the preconceptions of his predecessors. "The word `gene,' he wrote, "is completely free from any hypotheses."⁷ But it takes more than a new word to effect a complete break with the past. Weismann's determinants and de Vries' pangens were still the direct precursors of the gene, and inevitably some of the preconceptions underlying these earlier concepts carried over. Genes were hypothetical entities, but, like their precursors, they were particulate entities (Mendel himself had called his factors Elemente). Furthermore, whatever they were made of-indeed, even for those who thought of them as no more than a bookkeeping device-the capacity for faithful transmission from generation to generation remained built into the very notion, as it were, by definition.

No student of heredity, either before or after the watershed of 1900 thought of these hereditary elements literally as atoms, but the analogy with the fundamental units of physics and chemistry continued to lurk in people's minds. As E. B. Wilson wrote in 1923, "Even if considered only as working instruments . . . these conceptions have a practical value almost comparable to that of the atomic theory as employed in chemistry and physics."⁸ To the extent that the very notion of an atom implied stability, the analogy would have seemed especially apt for thinking about the immutability of hereditary elements. But even after 1901, when it had first been observed that the elements of physics and chemistry could themselves undergo spontaneous "transmutation," physicochemical elements continued to serve as models for the elements of heredity-perhaps, with this new possibility, as even better models. In fact, the occasional spontaneous transmutation of atoms that were nevertheless essentially stable served biologists well, for it opened a way to reconcile genetics with evolution. Hereditary elements too must sometimes change-indeed, it was precisely the occasional occurrence of such changes (or mutations) that made experimental genetics possible in the first place, for it was the tracking of mutations through generations that constituted the core method of classical genetics. In this sense, the advent of quantum mechanics might be said to have been fortuitous for biologists, especially for those who continued to look to physics as a model for their own fundamental units.

The geneticist H. J. Muller was one. In 1921 he wrote: "It is not physics alone which has its quantum theory. Biological evolution too has its quanta-these are the individual mutations."⁹ Five years earlier, while still a student, Muller had already noted "the curious similarity which exists between the main problems of physics and of biology." Furthermore, he argued, finding the means to influence mutation "might obviously place the process of evolution in our hands," just as the power to direct the transmutation of the elements could render "inanimate matter practically at our disposal." In conclusion, he proclaimed, "Mutation and Transmutation-the two keystones of our rainbow bridges to power!"¹⁰

Muller's contributions to the history of experimental genetics are legendary, but he was also a theoretician and a visionary. And in neither his theoretical nor his visionary writings did he ever lose sight of what remained, for him, the central question: Just what sort of entity is a gene? Nor was he ever able to provide an answer.

SCHROEDINGER' S QUESTION

The rise of classical genetics over the first half of this century is one of the great success stories of our time, and its history has been well documented. Yet despite its many successes, the question remained: What kind of object might a gene be that it can reproduce itself with such remarkable fidelity, generation after generation? Indeed, it was this very property of the gene, its manifestation of "a durability or permanence that borders upon the miraculous" that so mystified the physicist Erwin Schroedinger in the early 1940s as to inspire him to take on that grandest of all questions, "What is life?"¹¹ To Schroedinger, it seemed evident that the question of what endowed the gene with such durability, what lent it its apparent immunity to the second law of thermodynamics - with a "permanence unexplainable by classical physics"-got at the very core of the distinction between living and nonliving beings. He believed not only that the answer to this question would solve the problem of heredity but also that it would explain the equally remarkable capacity of organisms to maintain themselves against the ravages of entropy, to keep on going for so much longer than the laws of physics would lead us to expect. It would give us, in short, the secret of life.

Schroedinger, alas, did not find the secret of life. As one of the fathers of quantum mechanics, he not surprisingly sought the solution of this problem in the explanation that theory had already provided for the chemical stability of molecules. The particular model of gene structure on which he based his hope had been proposed in 1935 by two physicists and a geneticist.¹² In their picture, the gene was figured as a quantum mechanical system that derived its stability from the height of the energy barrier separating one state from another. The theoretical contributions to the model were made by Max Delbrück (a student of Niels Bohr), and accordingly Schroedinger referred to it as "Delbrück's model," adding to Delbrück's speculations the provocative proposal that the gene is not just a large molecule but an "aperiodic crystal or solid." Indeed, he saw "no alternative to the molecular explanation of the heredity substance." As he wrote, "The physical aspect leaves no other possibility to account for its permanence. If the Delbrück picture should fail, we should have to give up further attempts."¹³

Yet the Delbrück picture did fail, and with that failure so too did Schroedinger' s solution. Nonetheless, even with all its defects, the very effort of so prominent a physicist to solve so fundamental a biological problem served as powerful inspiration for an entire generation of young physicists and biologists, encouraging them in their own efforts to find the molecular structure of the gene. And soon they succeeded. Success, however, came not as a consequence of theoretical speculation but out of a series of experimental reports that narrowed the search to the structure of a specific chemical candidate.

The route by which biologists came to accept DNA as the genetic material has a long, rich, and well-documented history.¹⁴ In most popular accounts, however, that history begins with the paper by Avery, MacLeod, and McCarty which demonstrated through direct experiment that DNA was the carrier of biological specificity (at least in bacteria).¹⁵ This now-classic paper was published in the same year as Schroedinger's book. In it, the authors provided strong evidence arguing that DNA "must be regarded not merely as structurally important but as functionally active in determining the biochemical activities and specific characteristics of pneumococcal cells."¹⁶ But not everyone was immediately persuaded. Indeed, it was only after the almost equally famous "blender experiment" of Hershey and Chase in 1952 that most biologists were won over to the view that the genetic material was made up of DNA.¹⁷

Less than a year later, Watson and Crick struck gold. When the last piece of their model for the structure of DNA fell into place in the spring of 1953, Watson tells us that Crick "winged into the Eagle to tell everyone within hearing distance that we had found the secret of life."¹⁸  It is not hard to understand his enthusiasm. Not only did that structure provide a mechanism for the gene's remarkable capacity for self-replication-a mechanism that was stunning in its very simplicity-but also, and at the same time, it provided an (equally simple) explanation for the stability of the gene for the ostensibly miraculous fidelity with which it could be copied over so many generations. Complementary base-pairing could, at one fell swoop, do the work of both replication and conservation, or so it seemed (Figure 1).

If one assumed that DNA was an intrinsically stable molecule (as people did) and that complementary base-pairing proceeded without error, then nothing more would be required. In a sense, one might even say that Watson and Crick's triumph provided retrospective vindication of Schroedinger's own earlier speculations. From the vantage point of the simple picture that now emerged, his proposal of an aperiodic crystal or solid for the structure of the gene (and perhaps of the entire chromosome) acquired, at least in hindsight, an aura of prophecy.

Watson and Crick's achievement stands unrivaled in the annals of twentieth-century biology, and it is worth pausing for a moment to register the extraordinary sense of satisfaction that accompanied their findings. Since the beginning of the century, the notion of the gene as a self-replicating entity that carried the secret of its immortality in its very structure had been a staple of genetics, but no one had ever been able to say what kind of material such an entity might be made of. Now, after more than fifty years, an actual chemical substance-one already known to be a basic constituent of chromosomes-had been shown to have the necessary defining properties. Even before a mechanism was worked out by which the sequence of nucleotides in a DNA molecule could be translated into a sequence of amino acids in a protein molecule, confidence was widespread that the material basis of genetics had finally been established.

The decade that followed seemed little short of heroic. All the fundamental problems of biology yielded quickly, without difficulty or surprise. In 1968 an article appeared in Science entitled "That Was the Molecular Biology That Was." Here, Gunther Stent, an active participant in the exciting new research, described the approaching decline of the discipline that was "only yesterday an avant-garde but today definitely a workaday field."¹⁹ In Stent's view, by 1963 molecular biology had already entered what he called its "academic phase." He wrote, "All hope that paradoxes would still turn up in the study of heredity had been abandoned long ago, and what remained now was the need to iron out the details." ²⁰

IRONING OUT THE DETAILS

The history of science is replete with irony, and the aftermath of Watson and Crick's tour de force offers no exception. As everyone now knows, Stent could not have been more wrong. Molecular biology's course after 1968 was anything but a decline. Only two years later, with the isolation of a restriction enzyme that can recognize and cut DNA molecules at specific sites, the field was launched into a new, and in some ways even more productive, era. Restriction enzymes are the basis of the powerful techniques of recombinant DNA that have opened vast new vistas and, in doing so, have yielded so many surprises.

One such surprise bears directly on Weismann's original question, that is, on the source of genetic stability. To be sure, DNA is copied in living cells with a fidelity that borders on the miraculous. But contrary to expectations, the structure of DNA provides only the beginning of an explanation for this high fidelity. In fact, left to its own devices, DNA cannot even copy itself. DNA replication will simply not proceed in the absence of the enzymes required to carry out the process (Figure z). Moreover, DNA is not intrinsically stable: its integrity is maintained by a panoply of proteins involved in forestalling or repairing copying mistakes, spontaneous breakage, and other kinds of damage incurred in the process of replication. Without this elaborate system of monitoring, proofreading, and repair, replication might proceed, but it would proceed sloppily, accumulating far too many errors to be consistent with the observed stability of hereditary phenomena-current estimates are that one out of every hundred bases would be copied erroneously. With the help of this repair system, however, the frequency of mistakes is reduced to roughly one in 10 billion (Figure 3).²¹

In point of historical fact, however, indications that the cell was involved in the maintenance of genetic stability had begun to emerge well before the "recombinant DNA revolution," even if they failed to attract much attention. The first signs came in the late 1950s and early 1960s, from studies of radiation damage in bacteria and bacterial viruses (phages) at Oak Ridge National Laboratory, and especially from the discovery that certain kinds of damage could be spontaneously reversed. Bernard Strauss, who was a participant in this early work, writes, "The discovery that genic material did not stand permanently aloof from cellular metabolism was a major surprise of the 1950's and 1960's."²² Yet in the wider community of molecular biologists, the surprise of these new findings was slow to register and their implications were even slower to dawn.

Rollin Hotchkiss of the Rockefeller Institute was a prominent member of that wider community, but he was also something of an exception. As early as 1968, he wrote: "We are turning away from the DNA of a decade just over, a relatively unchanging, stable reservoir of linear information. It has had its convincing tellings and smugly one-dimensional retellings and become `well known' (which is to say, often mentioned). But it has become necessary to face the fact that DNA grows, issues directives, opens up, closes, twists, and untwists. We are coming to realize how marvelously communicative it is, and that it is not an aloof, metabolically inert material, but instead one maintained and exquisitely balanced in an actively supported status quo. "²³

A more typical response, however, is suggested by Franklin Stahl, one of the central figures in the heroic age of molecular biology. In a recent history of the discovery of DNA repair mechanisms, Errol Friedberg (another participant in the early repair field) reports Stahl's response when Friedberg "challenged him with the question of why the concept of gene/DNA repair was late in coming": "I suspect because of a widespread belief (unspoken I suspect, but amounting to worship) among geneticists that the genes are so precious that they must (somehow) be protected from biochemical insult, perhaps by being carefully wrapped. The possibility that the genes were dynamically stable, subject to the hurly-burly of both insult and clumsy (i.e., enzymatic) efforts to reverse the insults, was unthinkable." ²⁴

When I later queried him further, Stahl acknowledged that much of the early work on repair in radiation biology was overlooked.²⁵ But to the widespread belief that stability inhered in the gene itself he added two other factors: one disciplinary ("the investigators usually seemed to have little sense of genetics") and the other political (radiation biology was "somewhat suspect" because it was under the sponsorship of the Atomic Energy Commission). For Stahl, it was only after Evelyn Witkin and Miroslav Radman's work implicating recombination with repair (discussed below) that "we could no longer blow off investigations on repair." "Now, not only was the geneticists' domain of recombination invaded by the repair people but their domain of mutagenesis was also. The walls were not only breached, but they were toppled."²⁶

Over the last fifteen years the field of DNA repair has truly exploded. In 1994 the journal Science gave its "Molecule of the Year" award to the enzymatic repair machinery, and for good reason. The mechanisms already known to be involved in proofreading, editing, and repairing damaged or miscopied DNA can scarcely fail to astonish us-by their ingenuity their complexity, and perhaps especially by their implications for our understanding of evolution. But they are still far from clearly understood, and, as is inevitably the case with cutting-edge research, as yet subject to considerable debate. For all these reasons, the sketch that follows is simultaneously technical and provisional, and in both cases unavoidably so.

Three different kinds of processes seem to be involved in ensuring the fidelity of replication as it proceeds. The first works by helping to select the correct nucleotide for complementary binding. The second works by checking the most recently added nucleotide and immediately removing it if it should fail the test of complementarity. The third comes into action only after a new strand has been synthesized, and it works by repairing mismatches that might have occurred in spite of the first two error-avoidance mechanisms. A fourth set of repair mechanisms-first observed in the early work on photoreactivation-come into play later, in response to environmentally inflicted damage (caused, for example, by ultraviolet light). If the damage has been confined to a single strand, these excision repair mechanisms can reverse it with little chance of error by excising the damaged section and allowing it to be recopied from the undamaged strand (Figure 3).²⁷

The stability of gene structure thus appears not as a starting point but as an end-product-as the result of a highly orchestrated dynamic process requiring the participation of a large number of enzymes organized into complex metabolic networks that regulate and ensure both the stability of the DNA molecule and its fidelity in replication.²⁸ As the late Robert Haynes has written, "The stability of genes is now seen to be more a matter of biochemical dynamics, than of the molecular 'statics' of DNA structure. The genetic machinery of the cell provides the most striking example known of a highly reliable, dynamic system built from vulnerable and unreliable parts."²⁹

THE LIMITS OF GENETIC STABILITY

Even with such an elaborate process of proofreading and repair, genetic stability is not absolute, and fortunately not. If genes were truly immortal, and if their replication proceeded with perfect fidelity, the evolution of new genetic structures would never have been possible. As Darwin so clearly understood, change too is a desideratum of life, and the question naturally arises: How much genetic instability (or mutability) is necessary? How much would be required to accord with the pace at which evolution has actually occurred? Indeed, is it possible that the balance between genetic stability and mutability observed in the organisms we see today is itself a product of evolution? In other words, might selective pressures have operated on the very capacity to evolve, giving rise to the evolution of special mechanisms for generating change?

Evolution by natural selection depends on the occurrence of that rarest of events, mistakes that proved beneficial. But the fact remains that the vast majority of naturally occurring mistakes are either harmful or neutral.³⁰ Would it not therefore have been advantageous to the survival of cells and organisms to have developed mechanisms ensuring an even higher accuracy of replication, permitting fewer mistakes?

Given what is known, it is not hard to imagine ways of increasing genetic stability. But one of the most interesting insights to come out of work on repair mechanisms is the recognition that the advantage of increased fidelity in replication is not fixed but rather depends on both the organism and the conditions in which that organism finds itself. The critical dependence of genetic stability on proofreading and repair enzymes may have come as a great surprise, but more surprising yet was the discovery of "repair" mechanisms that sacrifice fidelity in order to ensure the continuation of the replication process itself-and hence the survival of the cell. Far from reducing error, such mechanisms actively generate variations in nucleotide sequence; moreover, it appears that when and where they come into play is itself under genetic regulation. With such findings as these, Barbara McClintock's remarks in her 1983 Nobel address describing the genome as "a highly sensitive organ of the cell, monitoring genomic activities and correcting common errors, sensing the unusual and unexpected events, and responding to them" no longer seem quite so far-fetched.³¹

The first indications of an error-prone mechanism of repair also came out of early studies of UV radiation damage in bacteria, and once again the implications were slow to dawn. Evelyn Witkin was a key figure in this work. In 1967, after isolating a mutant of E. coli in which this activity was repressed, she argued for a mechanism of stress-induced mutagenesis that might itself be under genetic control. Soon after at the Massachusetts Institute of Technology, Maurice Fox conducted a preliminary experiment showing that a closely related error-prone repair activity implicated in the reactivation of UV-irradiated bacterial viruses (and known as Weigle-reactivation) depends on the synthesis of new protein-an observation that led him to suggest to the young Miroslav Radman that one or more specific genes might be involved in the induction of error-prone repair.³² Four years later, Radman coined the expression "SOS response" to designate both this and Witkin's UV-induced error-prone (or hypermutagenic) modes of replication.³³

For Radman, son of a fisherman, the meaning of SOS was clear: "An international distress signal to save endangered life on the sea."³⁴ The term stuck and is now widely understood as referring to last-ditch efforts invoked to save the cell from going under. In his first publication on the subject, however, Radman was somewhat more circumspect: "Because of its `response' to DNA-damaging treatments," he wrote, "we call this hypothetical repair `SOS repair'... In order for SOS repair to function it should require specific genetic elements, the inducing signal and de novo protein synthesis."³⁵ The questions that needed answering were obvious. First, what are these specific genes and proteins? And, second, how are they induced? But without the tools with which to answer such questions, further progress proved desultory, and soon even Radman's attention turned to other problems.

A quarter of a century later, the regulation of mutability has become one of the hottest topics in molecular biology; and with the new analytic techniques that have now become available, many aspects of the biochemical machinery involved in such regulation have been elucidated. But with every step toward elucidation, the picture is rendered ever more complex by the increasing wealth of detail. Scores of proteins have already been implicated in the cell's response mechanisms, and reports of new players appear with every passing month.

Furthermore, as the picture becomes more and more complicated, so too does it appear ever more radical in its implications. To be sure, there is nothing radical in the notion that genetic stability and variability (or mutability) are complementary in their effects. Now, however, stability and mutability are proving to be flip sides of one another in the specific mechanisms by which they are controlled. Both are at the merry of enzymatic processes, and apparently equally so. Moreover, not only are the mechanisms controlling stability and mutability held in a delicate balance, but that very balance is under cellular regulation, and it shifts in response to the particular environment in which the cell finds itself.

All this is a far cry from the traditional view of DNA as an inherently stable molecule subject to occasional random errors, and it suggests an even further departure from the traditional view of evolution as a process of cumulative selection of those exceedingly rare mutations that happen to result in increased fitness. At least to many, the new picture seems to accord far better with McClintock's image of the genome as "a highly sensitive organ" than it does with the neo-Darwinian view of the genome as a strictly passive partner in the evolutionary two-step of variation and natural selection.

That regulation of genetic stability and mutability is a feature of all living systems is now widely accepted. The phenomenon has been most intensively studied in bacteria, and here a number of mechanisms that enhance mutation rates are at least partially understood. In many cases, increased mutation rate is associated with a defect in genes required for the overall maintenance of genetic stability. Such genes are sometimes called mutator genes, although the terminology is a bit misleading because mutation rates increase only when the gene is defective. For example, a mutation may eliminate or reduce the efficacy of an enzyme required for proofreading or error correction; it may interfere with the process by which newly synthesized strands of DNA can be distinguished from the old strands (for example, by methylation of the old strand); or it may eliminate or disable an enzyme involved in excision repair. Alternatively, a mutation might lead to increased mutation rates by disabling the regulatory mechanism that represses the activity of the error-producing SOS system under normal conditions-that is, in the absence of extensive (stress-induced) damage.

It is important to note, however, that the unmutated (or wild-type) SOS system becomes a generator of increased mutation rates when called into action by conditions of severe streSS.³⁶1n cultures of bacteria that have not undergone deregulating mutations (that is, in wild-type cultures), the SOS system is activated only by specific signals (generated, for example, by the persistent presence of single-stranded DNA or interruptions in the process of replication). Such signals activate a variety of mechanisms that allow the damaged region to be bypassed, filled in, or exchanged with a homologous region of DNA nearby (through recombination). Simultaneously, they lead to the inactivation of some of the normal proofreading functions, thus permitting replication to proceed even with the accumulation of many mistakes. Once replication has been completed, however, the SOS system returns to its normal repressed state and the machinery of proofreading and error-correction resumes its customary vigilance.

Most of these SOS functions are directly implicated in repair of one sort or another-in the sense, that is, that they make it possible for replication to proceed. But Radman and his colleagues now argue that the SOS system also functions merely to generate diversity, as if for its own sake, without serving any obvious repair function.³⁷ Perhaps he is right, but how could one possibly make evolutionary sense of such an idea?

THE EVOLUTION OF EVOLVABILITY-MOLECULAR BIOLOGY'S CHALLENGE TO NEO-DARWINISM

We now know that mechanisms for ensuring genetic stability are a product of evolution. Yet a surprising number of mutations in which at least some of these mechanisms are disabled have been found in bacteria living under natural conditions. Why do these mutants persist? Is it possible that they provide some selective advantage to the population as a whole? Might the persistence of some mutator genes in a population enhance the adaptability of that population?

Apparently so. New mathematical models of bacterial populations in variable environments confirm that, under such conditions, selection favors the fixation of some mutator alleles and, furthermore, that their presence accelerates the pace of evolution.³⁸ Recent laboratory studies of bacterial evolution provide further confirmation,³⁹ lending support to the notion that organisms have evolved mechanisms for their own "evolvability."⁴⁰

Mutator genes, however, are constitutive-that is, they give rise to high mutation rates even in the absence of provocation. Thus, over time they might enhance the adaptability of a population, but unlike inducible systems such as the SOS system, they offer no obvious adaptative advantages to individual organisms.⁴¹ Might there be a connection between these quite different mechanisms for generating rapid change? Radman believes there is. He writes: "Mutagenesis has traditionally been viewed as an unavoidable consequence of imperfections in the process of DNA replication and repair. But if diversity is essential to survival, and if mutagenesis is required to generate such diversity, perhaps mutagenesis has been positively selected for throughout evolution."⁴² In support of this view, he cites the recent identification of several enzymes involved in the SOS system. Such mutases, as Radman calls them, are "designed to generate mutations,"⁴³ And because they are inducible, they can be argued to enhance the resources of the organism as well as those of the population for coping with unanticipated environmental challenges. "Chance," as one of the organizers of a recent conference on "Molecular Strategies in Biological Evolution" puts it, "favors the prepared genome."⁴⁴

The notion that mechanisms for evolvability could themselves have evolved is a serious provocation for neo-Darwinian theory, for it carries the heretical implication that organisms provide not just the passive substrate of evolution but their own motors of change; it suggests that they have become equipped with a kind of agency in their own evolution. It also strongly implies the operation of selection on levels higher than the gene, and higher even than the individual organism. As James Shapiro writes, "These molecular insights lead to new concepts of how genomes are organized and reorganized, opening a range of possibilities for thinking about evolution. Rather than being restricted to contemplating a slow process depending on random (i.e., blind) genetic variation and gradual phenotypic change, we are now free to think in realistic molecular ways about rapid genome restructuring guided by biological feedback networks."⁴⁵

Shapiro, like Radman, studies bacteria. But the idea that organisms have evolved their own mechanisms for change has come to extend far beyond bacteria, and it has been gathering increasing support from biologists across a wide range of specialties. As far as I can tell, the first use of the expression "the evolution of evolvability" was as the title of a paper on "artificial life" by Richard Dawkins.⁴⁶ Dawkins wrote, "A title like `The Evolution of Evolvability' ought to be anathema to a dyed-in-the-wool, radical neo-Darwinian like me! Part of the reason it isn't is that I really have been led to think differently as a result of creating, and using, computer models of artificial life."⁴⁷ It was by tinkering with such computer simulations that Dawkins was led to hypothesize "a kind of higher-level selection, a selection not for survivability but for evolvability."⁴⁸ For many others, however, what led them to the notion was trying to make sense of their own and their colleagues' experimental observations. This is especially the case for the increasing number of developmental and evolutionary biologists seeking to bring some coherence to the accumulating mass of data attesting to both conservation and transformation of cellular and developmental systems across the evolutionary spectrum.

As the term is employed in these wider communities, evolvability refers to the capacity to generate any kind of heritable phenotypic variation upon which selection can act. It may be based on individual mutator genes or on higher-level genetic or epigenetic networks. The particular appeal of such an extended sense of the notion of evolvability lies in its ability to shed light on the evolution of developmental systems; Per Alberch may have been the first to make this point.⁴⁹ But probably the most extensive argument made to date for the role of evolvability in the evolution of complex organisms comes from John Gerhart and Marc Kirschner. As they wrote in their 1997 book, Cells, Embryos, and Evolution, "Throughout [the history of genetics], the organism remained a black box, translating random change in its genes into phenotypic variation to be acted on by selection. This black box is being quickly opened up by modern biology. In it we find that the connections between genotype and phenotype have been crafted by evolution to collaborate with evolution."⁵⁰

The particular concern of these authors is with the explosion of diversity in the evolution of metazoan organisms that appeared in the Cambrian period, and with the extensive conservation of core genetic, cellular, and developmental processes that accompanied this diversity. From their analysis of the interdependence of such diversification and conservation they draw this principal conclusion: "As we look at breakthroughs in metazoan design since the pre-Cambrian, they seem to involve a succession of new attributes of evolvability, as if evolvability has itself evolved"-as if the major transitions of evolution depended upon the acquisition of ever more sophisticated motors of change.⁵¹

WHAT IS LIFE?

How do all these findings affect our thinking about genetics, development, and evolution? Needless to say, their full implications have yet to be explored, but already at least two lessons can be inferred with reasonable confidence. The first concerns the nature of genetic stability and the second the dynamics of genetic transformation and hence of evolutionary change. Taken separately, each of these implies a major reversal of one of the key expectations with which the century began. Taken together, they imply radical modifications in our view of the relationships that tie genetics, development, and evolution together. First, the question of stability.

By now, we have abandoned the hope of finding in the molecular structure of particulate genes an adequate explanation for the stability of biological organization across generations. We have learned that genetic stability is itself a consequence of biological organization, and while it may

be a prerequisite for natural selection, the mechanisms for guaranteeing such stability are themselves an achievement of evolution. Furthermore, these mechanisms are not static but dynamic, and an explanation of how they do their job will have to be sought in the complex systems of cellular dynamics that are at one and the same time the products and the safeguard of genetic information.

What might such an explanation look like? One of the ironies of this history is that the very man who had been the source of the model leading Schroedinger astray gave us a clue. At a meeting in Paris in 1949 on mechanisms of genetic continuity-long after his quantum mechanical model of gene structure had failed-Max Delbrück sketched a quite different kind of model for achieving biological stability. As he explained, a system of cross-reacting and mutually inhibiting chemical reactions can lead to not just one stable steady state but to multiple steady states.⁵² With such a mechanism, stability does not depend on the immutability of individual particles but solely on the dynamics of their interaction. Delbrück introduced this new model as an explicit alternative to arguments that had been presented for the existence of cytoplasmic genes-that is, as a nongenetic way to account for the stability of certain kinds of cellular inheritance, and not as a way to account for genetic stability as such.⁵³ And indeed, the primary use to which this model has subsequently been put has been to stimulate work on the stable steady states of metabolic regulatory networks. Yet it can also be seen (as it was by some) as harking back to a pre-Weismannian tradition in which heredity in general was regarded as a manifestation of biological regulation-in the words of David Nanney, as a "type of homeostasis."⁵⁴

But homeostasis (or self-maintenance) has turned out to be only one part of the story. The capacity to generate rapid change in times of stress suggests a far more dynamic mode of organization than anything suggested by Delbrück's model of steady states. Yet even so, that simple model may point us in the right direction, and perhaps especially so for thinking about Schroedinger's question, "What is life?" Historically, biologists have in fact been quite divided in their approach to an answer. Throughout Weismann's own century, the dominant tradition held that the defining feature of life was its organization. But after Weismann, a different tradition came to dominate, one which tended to define life in terms of genes and their replication. Still, a minority tradition has continued to look to self-maintaining (or autocatalytic) metabolic systems for the essence of life.

This divide is perhaps most conspicuous in debates about the origin of life. Which came first-genes or cells? Replication or self-maintenance? In 1985 Freeman Dyson, another physicist, published a small book called origins of Life. Revisiting Schroedinger's question, he suggests that Schroedinger's approach exemplified a long-standing over-preoccupation with genes. Life, he argues, requires not just nucleic acid but also a metabolic system for self-maintenance; hence, the overwhelming likelihood is that it had not one but two origins. The emergence of living systems as we know them could have come about as a result of a symbiotic fusion between two independently evolved prior subsystems-one a rapidly changing set of self-reproducing but error-prone nucleic acid molecules and the other a more conservative autocatalitic metabolic system specializing in self-maintenance.

Over the last fifteen years, Dyson's picture of dual origins of life has gained increasing currency. According to this hypothesis, out of the interactions permitted by the conjoining of these two subsystems emerged a mechanism of heredity able to ensure the presence of a particular genotype over a period of time long enough for natural selection to begin its work and yet flexible enough to generate the variability on which natural selection could act. For then and only then would it be possible to evolve such exquisitely creative mechanisms as are needed for the genetic code, individuality, multicellularity, or sex. The rest, as they say, is history. But what a dynamic history it would have had to be.

NOTES

1. Darwin, C. (1859), p. 167.

2. Weismann, A. (1885), quoted in Gabriel, M. L., and Fogel, S. (1955) p. 200.

3. De Vries, H. (1889 [19101), p. 13.

4. Quoted in Portugal, F. H., and Cohen, J. S. (1977), P. toy. For a particularly interesting discussion of Weismann in relation to the increasing distance over the course of the century between genes and the body, see Griesemer, J. R. (forthcoming).

5. Wilson, E. B. (1896), p. 13.

6. Quoted in Portugal, F. H., and Cohen, J. S. (1977), P. too.

7. Johannsen, W. (1909), P. 124.

8. Wilson, E. B. (1923), P. 280.

9. Quoted in Carlson, E. A. (1971), p. 161. To help frame Muller's thinking, a few historical markers might be helpful: 1896, radio activity discovered; 1900, rediscovery of Mendelian "factors" (particulate genetic elements); 1901, spontaneous "transmutation of elements" observed by Rutherford and Soddy (so named by Soddy in 1901); 1902, identification of nuclear chromosomes as the site of genetic factors; 1909, coinage of the term gene; 1911, Rutherford's discovery of the atomic nucleus; 1919, Rutherford induces first "artificial transmutation." to 

10. Quoted in Carlson, E. A. (1971), P. 161. Indeed, after Rutherford's success in 1919 in inducing a transmutation of the elements, Muller pursued his own search for a means of inducing mutation with that precedent directly in mind, even entitling his discovery of X-ray induced mutations "Artificial Transmutation of the Gene" (Muller, 1927).

11. Schroedinger, E. (1944), P. 49.

12. Timoféeff-Ressovsky, N. W., Zimmer, K. G., and Delbrück, M. (1935)

13. Schroedinger, E. (1944), p. 61.

14. One of the best accounts (and surely the most extensively documented discussion) of this history is to be found in Olby, R (1974)

15. Avery, O. T, MacLeod, C. M., and McCarty M. (1944)

16. Ibid., P. 155.

17. Hershey, A. D., and Chase, M. (1952). It was the use of a Waring blender to separate the viruses from their host bacteria that lent this experiment its name.

18. Watson, J. D. (1968), P. 197.

19. Scent, G. (1968), p. 390.

20. Ibid., p. 394.

21. See, for example, Radman, M. (1988).

22. Strauss, B. (1995), P. 1511.

23. Hotchkiss, R (1968), P. 857.

24. Letter to Friedberg, E., February 21,1995, quoted in Friedberg, E. (1997), P. 17.

25. Frank Stahl to E. F. Keller, September 1,1997 (e-mail).

26. In response to the question of whether people were surprised, Stahl wrote, "I think (for me) the realization dawned rather slowly, precluding any sense of surprise."

27. Additional mechanisms guaranteeing the integrity of the entire chromosome have also been identified, but these will not be discussed here.

28. To complicate matters even further, evidence is now beginning to emerge suggesting that other enzymes, organized into ocher repair pathways, may work to monitor and correct errors in transcription, translation, and even in protein structure. Such additional mechanisms would ensure a degree of stability of biological organization going well beyond that of merely genetic stability.

29. Haynes, R (1988), P. 577.

30. In bacteria, the ratio of harmful to adaptive mutations is estimated at 100,000 to 1.

31. McClintock, B. (1983). Reprinted in Federoff, N., and Botstein, D. (1992).

32. Interview with M. S. Fox by Leslie Barber, November 1g, r998. Although the initial experiment was never published, an improved version, performed in Radman's lab, was reported some years later (Defais, M., Caillet-Fauquet, P., Fox, M. S., and Radman, M.,1976).

33. See Witkin, E. (1989), p. 32.

34. Quoted in Friedberg, E. (1997), p. 277.

35. Radman, M. (7974) P. 134. Today, however, Radman interprets this early proposal more radically-chat is, as predicting the existence of certain enzymes that would specifically enhance diversity and adaptability in the endangered population by allowing individual cells to mutate when their survival is under threat; see Radman, M. (7999), p. 866.

36. Indeed, bacterial mutants that lack key components of the SOS system fail to show the usual increase in mutagenesis expected under exposure to ultraviolet light.

37. Taddei, F., Vulic, M., Radman, M., and Matic, I. (7997)

38. Taddei, F., Radman, M., Maynard Smith, J., Toupance, B., Gouyon, P. H., and Godell, B. (1997)

39. Sniegowski, P. D., et al. (1997).

40. Radman, M., Matic, I., and Taddei, I. (1999)

41. The distinction between organism and population is blurred, however, by the fact that the effect of such inducible (or transient) mutagenesis is to alter the organism. Thus here coo the evolution of such mechanisms requires the evocation of some kind of group selection, operating either on the level of the population of genes in a given genome or on the level of a population of organisms.

42. Radman, M. (1999) P. 866.

43. Ibid.

44. Caporale, L. H. (1999).

45. Shapiro, J. A. (1999) P. 32.

46. Dawkins, R (1988).

47. Ibid., p. 201.

48. Ibid., p. 218.

49. Alberch, P. (1991).

50. Gerhart, J., and Kirschner, M. (1997), p. 613.

51. Kirschner, M., and Gerhart, J. (1998), p. 8427.

52. Delbrück, M. (1949).

53. For further remarks on the stability of cellular inheritance, see Chapter 3, n. 32.

54. Nanney, D. L. (1957), P. 134.
    
    
    

REFERENCES

Abelson, H., Allen, D., Coore, D., Hanson, C. P., Homsy, G., Knight, T. F., Jr., Nagpal, R, Rauch, E., Sussman, G. J., and Weiss, R 1999. Amorphous computing. White Paper, MIT. [local HTML file 9kb]  [http://www.swiss.ai.mit.edu/projects/amorphous/workshop-sept-99/hal-tk-gjs.pdf 16kb]

Alberch, P. 1991. From genes to phenotype: dynamical systems and evolvability. Genetica 84:5-11.

Alberts, B., Bray, D., Lewis, J., Raff, M., Roberts, K., and Watson, J. D. 7994 Molecular Biology of the Cell. New York: Garland.

Angier, N. 1993. When a vital gene is missing, understudies fill in. New York Times, Sept. 7,1993, P- C3

Apter, M. J. 1966. Cybernetics and Development. Oxford: Pergamon.

Apter, M. J., and Wolpert, L. 1965. Cybernetics and development. J. Theor. Biol, 8:244-257.

Atlan, H., and Koppel, M. 19go. The cellular computer DNA: program or data. Bulletin of Math. 52(3):335-348.

Avery, O. T., MacLeod, C. M., and McCarty, M. 1944. Studies on the chemical transformation of pneumococcal types. J Exp. Med. 79:137-158.

Avital E., and Jablonka E. 2000. Animal Traditions: Behavioural Inheritance in Evolution. Cambridge: Cambridge University Press.

Bailey, J. E. 1999. Lessons from metabolic engineering for functional genomics and drug discovery. Nature Biotech. 17:616-618.

Baltimore, D. 1984. The brain of a cell. Science 84:150.

Bateson, W. 1906. The progress of genetic research. In Third Conference on Hybridization and Plant Breeding, pp. 90-97. London.

Beadle, G. W. 1945. The genetic control of biochemical reactions. Harvey Lectures 40:179-194.

Beadle, G. W., and Tatum, E. L. 1941. Genetic control of biochemical reactions in Neurospora. Proc Natl. Acad. Sci. 21:499-506.

Beermann, W. 1956. Nuclear differentiation and functional morphology of chromosomes. Cold Spring Harbor Symp.Quant. Biol. 21:217-232.

Birman, K. P., and van Renesse, R 1996. Software for reliable networks. Scientific American, May, pp. 48-51.

Black, D. L. 1998. Splicing in the inner ear: a familiar tune, but what are the instruments? Neuron 20:165-168.

Bonner, J. 1965. The Molecular Biology of Development. Oxford: Oxford University Press.

Braus, H. 1906. Ist die Bildung des Skeletts von den Muskelanlagen abhängig? Morph. Jahrb. 35:38-119.

Brenner, S., Dove, W, Herskowitz, L, and Thomas, R. 1990. Genes and development: molecular and logical themes. Genetics 126:479-486.

Briggs, R, and King, T. J. 1952. Transplantation of living nuclei from blastula cells into enucleated frogs' eggs. Proc. Natl. Acad. Sci. 38455-463.

Brooks, R A. 1990. Elephants don't play chess. Robotics and Autonomous Systems 6:3-13.

-------1991. New approaches to robotics. Science 253:1227-1232.

-------1999 Cambrian Intelligence. Cambridge: MIT Press.

Buguski, M. S. 1999. Biosequence exegesis. Science 286:453-455

Burian, R M. 1985. On conceptual change in biology: the case of the gene. In Evolution at a Crossroads: The New Biology and the New Philosophy of Science, ed. D. J. Depew and B. H. Weber, pp. 21-42. Cambridge: MIT Press.

Burtis, K. C., and Hawley, R S. 1999. The millennium flies in. Nature 401:125-126.

Campbell, K. 1999. Nuclear transfer in farm animal species. Sem. Cell & Dev. Biology 10:245-253.

Canguilhem, G. 1994. A Vital Rationalist. Cambridge: MIT Press.

Caporale, L. H. 1999. Chance favors the prepared genome. In Molecular Strategies in Biological Evolution, ed. L. H. Caporale. New York: New York Academy of Sciences.

Carlson, E. A. 1971. An unacknowledged founding of molecular biology: H. J. Muller's contributions to gene theory. J. Hist. Biology 4:149-170.

Coen, E. 1999. The Art of Genes: How Organisms Make Themselves. Oxford: Oxford University Press.

Collingwood, R G. 1940. An Essay on Metaphysics. Oxford: Clarendon Press.

Creager, A. N. H., and Gaudillière, J. P. 1996. Meanings in search of experiments and vice-versa: the invention of allosteric regulation in Paris and Berkeley, 19591968. Historical Studies in the Physical and Biological Sciences 27:1-89.

Crick, F. 1957. On protein synthesis. Symp. Soc. Exp. Biol. 12:138-163.

Cubas, P., V. C., and Coen, E. 1999. An epigenetic mutation responsible for natural variation in floral symmetry. Nature 410:157-161.

Darwin, C. 1859. On the Origin of Species. Fac. ed. Cambridge: Harvard University Press, 1964.

------1862. On the Various Contrivances by Which British and Foreign Orchids Are Fertilised by Insects, and on the Good Effects of Intercrossing. London: John Murray.

Dawkins, R 1988. The evolution of evolvability. In Artificial Life, ed. C. G. Langton, pp. 201-220. New York: Addison Wesley.

De Chadarevian, S. 1998. Of worms and programmes: "Caenorhabditis elegans" and the study of development. Studies in Hut. and Phil. of Biol. and Biomed. Sci. 29:81-105.

De Vries, H. 1889 [1910]. Intracellular Pangenesis. Chicago: Open Court.

Defais, M., Caillet-Fauquet, P., Fox, M. S., and Radman, M. 1976. Induction kinetics of mutagenic DNA repair activity in E. coli following ultraviolet irradiation. Molec. Gen. Genet. 148:125-130.

Delbrück, M.1948. A physicist looks at biology. In Phage and the Origins of Molecular Biology, ed. J. Cairns et al., pp. 9-22. Cold Spring Harbor: Cold Spring Harbor Laboratory Press.

------1949. Discussion. In Unités Biologiques Douées de Continuité Génétique. Lyon: Editions du CNRS, 33. (Trans. by Delbrück, ms, sent to Leo Szilard, Feb. 1, 1960.)

Duboule, D. 1997. The evolution of genomics. Science 279:555

Dyson, F. 1985. Origins of Life. New York: Cambridge University Press.

Friedberg, E. 1997. Correcting the Blueprint of Life: An Historical Account of DNA Repair Mechanisms. Cold Spring Harbor: Cold Spring Harbor Laboratory Press.

Gabriel, M. L., and Fogel, S., eds. 1955. Great Experiments in Biology. New York: Prentice-Hall.

Garcia-Bellido, A. 1998. Discussion of S. Brenner, "Biological Computation." In The Limits of Reductionism, ed. G. R. Bock and J. A. Goode, pp. 106-116. Novartis Symposium. Chichester: John Wiley.

Gaudillière, J. P. 1988. Un code moléculaire pour la différentiation cellulaire: la controverse sur les transferts d'ARN informationnel (1955-1973) et les étapes de diffusion du paradigme de la biologie moléculaire. Fundamenta Scientiae 9:429-467.

------1993 Molecular biology in the French tradition? Redefining local traditions and disciplinary patterns. J Hist Biol. 26:473-498.

Gehring, W. 1998. Master Control Genes in Development and Evolution. New Haven: Yale University Press.

Gelbart, W. 1998. Data bases in genomic research. Science 282:660.

Gerhart, J., and Kirschner, M. 1997. Cells, Embryos, and Evolution. Oxford: Blackwell Scientific.

Gilbert, S. F. 1991. Induction and the origins of developmental genetics. In A Conceptual History of Modern Embryology, ed. S. F. Gilbert, pp. 18I-2o6. New York: Plenum.

Gilbert, W 1978. Why genes in pieces? Nature 271:501.

------1992. Vision of the grail. In The Code of Codes, ed. D. J. Kevles and L. Hood, pp. 83-97.

Goldfarb, D. S. 1997. Whose finger is on the switch? Science 276:1814-1817.

Goodwin, B. 1985. What are the causes of morphogenesis? BioEssays 3(7)32-36.

Goodwin, B., Kauffman, S., and Murray, J. D. 1993. Is morphogenesis an intrinsically robust process? J. Theor. Biol. 163:135-144

Gould, S. J. 1989. Wonderful Life: The Burgess Shale and the Nature of History. New York: Norton.

Griesemer, J. R. 2000. Reproduction and the reduction of genetics. In The Concept of the Gene in Development and Evolution: Historical and Epistemological Perspectives, ed. P. Beurton, R Falk, and H-J. Rheinberger. Studies in Philosophy and Biology. Cambridge: Cambridge University Press.

Griesemer, J. R Forthcoming. The informational gene and the substantial body: on the generalization of evolutionary theory by abstraction. In Varieties of Idealization, ed. N. Cartwright and M. Jones. Poznan Studies. Amsterdam: Rodopi.

Griffiths, P. E., and Neumann-Held, E. M. 1999. The many faces of the gene. Bioscience 49(8):656-662.

Gurdon, J., Laskey, R. A., De Robertis, E. M., and Partington, G. A. 1979. Reprogramming of transplanted nuclei in amphibia. In International Review of Cytology Supplement 9 Nuclear Transplantation, ed. J. F. Danielli and M. A. DiBernardino, pp. 161-178. New York: Academic Press.

Haldane, J. S. 1931. The Philosophical Basis of Biology. Garden City: Doubleday, Doran.

Halder, G., Callaerts, P., and Gehring, W. 1995. Induction of ectopic eyes by targeted expression of the eyeless gene in Drosophila. Science 267:1788-1792.

Hall, Alan. 19go. Ras and GAP: who's controlling whom? Cell 61:921-923.

Hartwell, L. H., Hopfield, J. J., Leibler, S., and Murray, A. W. 1999. From molecular biology to modular cell biology. Nature Suppl. 402:C47.

Haynes, R 1988. Biological context of DNA repair. In Mechanisms and Consequences of DNA Damage Processing, pp. 577-584. UCLA symposium held at Taos, N.M., January 24-30. New York: Alan R Liss.

Heath, J. R, Kuekes, P. J., Snider, G. S., and Williams, R S. 1998. A defect-tolerant computer architecture: opportunities for nanotechnology. Science 280:1716-1721.

Hershey, A. D., and Chase, M. 1952. Independent functions of viral proteins and nucleic acid in growth of bacteriophage. J. Gen. Physiology 36:39-56.

Hieter, P., and Boguski, M.1997. Functional genomics: it's all how you read it. Science 278:601-602.

Holland, P. W. H. 1999. The future of evolutionary developmental biology. Nature Suppl. 402:C41.

Horner, J. R, and Dobb, E. 1997. Dinosaur Lives. New York: Harcourt Brace.

Horwich, P., ed. 1994. World Changes: Thomas Kuhn and the Nature of Science. Cambridge: MIT Press.

Hotchkiss, R. 1968. Metabolism and growth of gene substance. Cold Spring Harbor Symp. Quant. Biol. 33:857-870.

Jablonka, E., and Lamb, M. 1995. Epigenetic Inheritance and Evolution. New York: Oxford University Press.

Jacob, F. 1982. The Possible and the Actual. New York: Pantheon.

Jacob, F., and Monod, J. 1961a. On the regulation of gene activity. Cold Spring Harbor Symp. Quant. Biol. 26:193-err.

-------1961b. Genetic regulatory mechanisms in the synthesis of proteins. J. Molec. Biol. 3:318-356.

Jeffery, C. J. 1999. Moonlighting proteins [talking point]. Trends in Biochem. Sci. 24(1):8-rr.

Johannsen, W. 1909. Elemente der Exakten Erblichkeitslehre. Jena: Gustav Fischer.

-------1911. The genotype conception of heredity. Am. Nat. 45:129-159

Kant, I. 1790. Critique of Judgement. Rpt. Great Books 39:461-475, trans. J. C. Meredith. Chicago: Encyl. Brittanica, 1993.

Keller, E. F. 1983. A Feeling for the Organism. New York: Freeman.

-------1995. Refiguring Life: Metaphors of Twentieth-Century Biology. New York: Columbia University Press.

-------2000a. Decoding the genetic program. In The Concept of the Gene in Development and Evolution: Historical and Epistemological Perspectives, ed. P. Beurton, R Falk, and H -J, Rheinberger. Studies in Philosophy and Biology. Cambridge: Cambridge University Press.

-------2000b. Marrying the pre-modern to the post-Modern: computers and organisms after WWII. In Growing Explanations, ed. N. Wise.

-------2000c. Is there an organism in this text? In Controlling Our Destinies: Historical, Philosophical, Ethical, and Theological Perspectives on the Human Genome Project, ed. P. R Sloan. Notre Dame: University of Notre Dame Press, pp. 273-290.

Keller, E. F., and Ahouse, J. C. 1997. Writing and reading about "Dolly." Bioessays 19(8):741-742.

Kevles, D. J., and Hood, L., eds. 1992. The Code of Codes. Cambridge: Harvard University Press.

Kirschner, M., and Gerhart, J. 1998. Evolvability. Proc. Nat. Acad. Sci. 95:8420-8427.

Lander, E. 1996. The new genomics: global views of biology. Science 274:536-539.

Lewontin, R C. 1997. The confusion over cloning. New York Review of Books, Oct. 23.

Maes, P., ed. 1991. Designing Autonomous Agents. Cambridge: MIT Press.

Maeshiro, T., and Kimura, M. 1998. The role of robustness and changeability on the origin and evolution of genetic codes. Proc. Natl. Acad. Sci. 95:5088-5093

Mann, T 1924 [1952]. The Magic Mountain. New York: Knopf

Mayr, E. 1959. Where are we? Cold Spring Harbor Symp. Quant Biol. 24:114.

1961. Cause and effect in biology. Science 134:1501-1506.

Miklos, G. L. G., and Rubin, G. M. 1996. The role of the genome project in determining gene function: insights from model organisms. Cell 86:521-529.

Monod, J., Changeux, J. P., and Jacob, F. 1963. Allosteric proteins and cellular control systems. J. Mol. Biol. 6:306-329.

Monod, J., and Jacob, F. 1961. General conclusions: teleonomic mechanisms in cellular metabolism, growth, and differentiation. Cold Spring Harbor Symp. Quant Biol. 26:389-401.

Morange, M. 1998. La Part des Gènes. Paris: Editions Odile Jacob.

Morgan, H. D., Sutherland H. G. E., Martin, D. 1. K., and Whitelaw, E. 1999. Epigenetic inheritance at the agouti locus in the mouse. Nature Genetics 23(3):314-318.

Morgan, T. H. 1928. The Theory of the Gene. New Haven: Yale University Press.

-------1933. The relation of genetics to physiology and medicine. In Nobel Lectures . . . Physiology and Medicine, 1922-1941, pp. 313-328. Amsterdam, 1963.

-------1934. Embryology and Genetics. New York: Columbia University Press.

Muller, H. J. 1927. Artificial transmutation of the gene. Science 66:8487.

-------1929. The gene as the basis of life. First presented before the International Congress of Plant Sciences, Section of Genetics, Symposium on "The Gene," Ithaca, N. Y., August 19, 1926; published in Proceedings of the International Congress of Plant Science 1:897-921.

-------1936. Physics in the attack on the fundamental problems of genetics. Scientific Monthly 44:210-214.

-------1951. The development of the gene theory. In Genetics in the Twentieth Century, ed. L. C. Dunn. New York: Macmillan.

Nanney, D. L.1957. The role of the cytoplasm in heredity. In The Chemical Basis of Heredity, ed. W. D. McElroy and B. Glass, Pp. 134-163. Baltimore: Johns Hopkins Press.

Olby, R 1974. The Path to the Double Helix. Seattle: University of Washington Press.

Pattee, H. 1969. How does a molecule become a message? Dev. Biol. Suppl. 3:116.

Portin, P.1993. The concept of the gene: short history and present status. Quart. Rev. Biol. 68:173-223.

Portugal, F. H., and Cohen, J. S. 1977. A Century of DNA. Cambridge: MIT Press.

Radman, M. 1973. Phenomenology of an inducible mutagenic DNA repair pathway in Escherichia coli: SOS repair hypothesis. In Molecular and Environmental Aspects of Mutagenesis ed. L. Prakash et al., pp. 128-142. Springfield: Charles C. Thomas.

-------1988. The high fidelity of DNA duplication. Scientific American, August, PP. 4046.

-------1999. Mutation: enzymes of evolutionary change. Nature 401:866869.

Radman, M., Matic, L, and Taddei, F. 1999 The evolution of evolvability. In Molecular Strategies in Biological Evolution, ed. L. H. Caparole. Annals of NY Acad. of Science 870:146-155.

Rheinberger, H -J. 1995. Genes: a disunified view from the perspective of molecular biology. Gene Concepts and Evolution, MPIWG preprint 18.

-------1996. Gene concepts: fragments from the perspective of molecular biology. Gene Concepts in Development and Evolution, MPIWG preprint 123.

Rieger, R, Michaelis, A., and Green, M. M. 1968. A Glossary of Genetics and Cytogenetics. New York: Springer-Verlag.

Schiller, J. 1978. La notion d'organisation dans l'histoire de la biologie. Paris: Maloine.

Schroedinger, E. 1944. What Is Life? Cambridge: Cambridge University Press.

Shapiro, J. A. 1999. Genome system architecture and natural genetic engineering. In Molecular Strategies in Biological Evolution, ed. L. H. Caparole. Annals of the NY, Acad. Sci. 870:23-25.

Sniegowski, P. D., Gerrish, P. J., and Lenski, R E. 1997. Evolution of high mutation rates in experimental populations of E. coli. Nature 387:703-705.

Spemann, H. 1938. Embryonic Development and Induction. New Haven: Yale University Press.

Stein, L. A. Forthcoming. Interactive Programming, New York: Morgan Kaufmann Publishers.

Stephens, C. 1998. Bacterial sporulation: a question of commitment? Cum. Biol. (B44) 8(2):845-48.

Sterelny, K., and Griffiths, P. E. 1999. Sex and Death: An introduction to the Philosophy of Biology. Chicago: University of Chicago Press.

Stewart, C. 1997. Nuclear transplantation: an udder way of making lambs. Nature 385:769.

Strauss, B. 1995. Molecular pathologies. Science 270:1511-1513.

Strohmann, R. C. 1997. The coming Kuhnian revolution in biology. Nature Biotechnology 15:194-200.

Sturtevant, A. H.1932. The use of mosaics in the study of the developmental effects of genes. Proceedings of the Sixth Int. Cong. of Genetics, p. 304.

Sussman, G. J. 1999. Robust design through diversity. http://www.swiss.ai.mit.edu/projects/amorphous/workshop-sept-99/robust-diversity.pdf 

Taddei, F., Radman, M., Maynard Smith, J., Toupance, B., Gouyon, P. H., and Godell, B. 1997. Role of mutator alleles in adaptive evolution. Nature 387:700-702.

Taddei, F., Vuliæ, M., Radman, M., and Matiæ, 1. 1997. Genetic variability and adaptation to stress. In Environmental Stress, Adaptation and Evolution, ed. R. Bijlsma and V. Loescheke. Basel: Birkhaüser.

Tautz, D. 1992. Redundancies, development and the flow of information. BioEssays 14(4):26;-266.

Thomas, J. H. 199;. Thinking about genetic redundancy. Trends in Genetics 9(11):395-399

Timoféeff-Ressovsky N. W., Zimmer, K. G., and Delbrück, M. 1935. Über die Natur der Benmutation and der Genstruktur. Nachr. Ges. Wiss. Göttingen, math-phys. Kl. Fachgr. G(1):189-245.

Troland, L. T 1917. Biological enigmas and the theory of enzyme action. Am. Nat. 51:321-350.

von Neumann, J. 1956. Probabalistic logics and the synthesis of reliable organisms from unreliable components. In Automata Studies, ed. C. E. Shannon and J. McCarthy. Princeton: Princeton University Press.

-------1966. Theory of Self-Reproducing Automata, ed. A. W. Burks. Urbana; University of Illinois Press.

Waddington, C. H. 1942. Canalization of development and the inheritance of acquired characters. Nature 150:563-565.

-------1948. The genetic control of development. Symp. Soc. Exp. Biol. 2:145-154, New York: Academic Press.

-------1954 The cell physiology of early development. In Recent Developments in Cell physiology, ed. J. A. Kitching. London: Butterworth.

-------1957. Strategy of the Genes, London: Allen & Unwin.

-------1962. New Patterns in Genetics and Development. New York: Columbia University Press.

-------1968-1972. Towards a Theoretical Biology: An IUBS Symposium, pp. 14. Edinburgh: Edinburgh University Press.

-------1971. Biology, Purpose and Ethics. Worcester, MA: Clark University Press.

Wade, N. 1998. Animal's Genetic Program Decoded, in a Science First. The New York Times, p. 1.

Watson, J. D. 1968. The Double Helix. New York: Atheneum.

-------1992. A personal view of the project. In The Code of Codes, ed. D. J. Kevles, and L. Hood, pp. 164-173.

Watson, J. D., and Crick, F. 1953a. A structure for deoxyribose nucleic acid. Nature 171:737-738.

-------1953b. Genetical Implications of the Structure of Deoxyribonucleic Acid. Nature 171:964-967.

Weatherall, D. J. 1998. How much has genetics helped? Times Literary Supplement, January 30, pp. 4-5.

Weaver, W. 1949. Problems of organized complexity. American Scientist 36:143-156.

Weinberg, R A. 1985. The molecules of life. Scientific American 253(4)=4857

Weismann, A. 1885 [1889]. Continuity of the Germ Plasm. In Essays upon heredity and kindred biological problems, ed. E. Poulton et al. Oxford: Clarendon Press.

Wilkins, A. 1997. Canalization: a molecular genetic perspective. BioEssays 19(3):257-262.

Wilmut, I., Schnieke, A. E., McWhir, J., Kind, A. J., and Campbell, K. H. S. 1997. Viable offspring derived from fetal and adult mammalian cells. Nature 385:81o-813.

Wilson, E. B. 1896. The Cell in Heredity and Development. New York: Macmillan.

-------1923. The physical basis of life. Science 42(1471):277-286.

Wimsatt, W. C. 198t. Robustness, reliability and overdetermination. In Scientific Inquiry and the Social Sciences, ed. M. Brewer and B. Collins, pp. 124163. San Francisco: Jossey-Bass.

-------2000. Re-Engineering Philosophy for Limited Beings: Piecewise Approximations to Reality. Cambridge: Harvard University Press, forthcoming.

Witkin, E. 1989. Ultraviolet mutagenesis and the SOS response in Escherischia coli: a personal perspective. Environmental and Molecular Mutagenesis 14(S16):30-34.

Wolpert, L., and Lewis, J. H. 1975. Towards a theory of development. Federation Proceedings 34(1):14-20.

Yoxen, E.1986. Form and strategy in biology: reflections on the career of C. H. Waddington. In A History of Embryology: The Eighth Symposium of the British Society for Developmental Biology, ed. T. J. Horder, J. A. Witkowski, and C. C. Wylie. Cambridge: Cambridge University Press.

Yuh, C-H., Bolouri, H., and Davidson, E. H. 1998. Genomic cis-regulatory logic: experimental and computational analysis of a sea urchin gene. Science 279:1896-1902.