Enzymatic Adaptation and the Entrance of Molecular Biology into Embryology

Scott F. Gilbert

Published in Sarkar, S., History of Molecular Biology, Kluwer Press, Dordrecht. Permission obtained courtesy of Dr. S. Sarkar.

I. Introduction: The Problems of Synthesis

There are many problems whenever a synthesis is attempted between two divergent disciplines. This is especially true when the two disciplines have an historical enmity and when the assumptions and axioms of the disciplines are at opposite ends of the continuum that characterizes the sciences. Genetics and embryology are two such divergent disciplines that are presently being united through molecular biology, and there are great differences between the genetic and embryological sciences. Moreover, to effect this integration, genetics and embryology are being placed into the common language of molecular biology. The molecularization of embryology has come about gradually and grudgingly. It is certainly far from complete, and it may never be completed. Phenotype-level embryologists (those concerned primarily with developmental anatomy and morphogenesis) fear for the integrity of embryology as a discipline and fear the lack of funding if they don't construct cDNA libraries of their favorite organs (see Malacinski, 1991; Gilbert, 1992). They see the handwriting on the wall, and it is full of As, Cs, Gs, and Ts. The molecularization of embryology has come about from several sources: cancer research, studies of globin synthesis, the isolation of ribosomal RNA genes, and analysis of embryonic lethal mutations being some obvious paths. This essay seeks to look at some of the reasons for this delay in the molecularization of embryology and to see how research into bacterial metabolism enabled molecular biology to establish a major foothold in embryology.

The globin studies are extremely important for integrating transmission genetics, population genetics, molecular genetics, medical and human genetics, and developmental genetics into a coherent field. (Only behavioral genetics were left out). Both the thalassemias and the globin switching were seen as important analogues to embryonic cell differentiation. These became even more important after the operon concept had introduced the notion of regulator gene sequences into embryology. The expansion of embryology into developmental biology (which permitted the data on adult cells red blood cell precursors and lymphocytes to model embryonic development) also allowed the globin studies to become used for embryonic differentiation.

There are at least four major dimensions to any disciplinary synthesis: conceptual, professional, philosophical, and aesthetic. All these dimensions are being negotiated in the attempts to synthesize genetics with embryology, and this has caused many embryologists to worry that the result won't be so much a synthesis as a take-over of embryology by the powerful and already molecularized geneticists. The first and most obvious dimension is the conceptual dimension. In this case, the problem concerns: What is the relationship between the genotype and the organismal phenotype? How does the DNA of the cell nuclei generate the living, eating, mating organism? Our models of this relationship have changed dramatically over the past fifty years. In the 1930s, many embryologists thought that the genes merely put the finishing touches on the organism, but that cytoplasmic proteins controlled the major events of ontogenesis. Others (such as E. E. Just) were convinced that there was no relationship between the genes and the organismal phenotype. The nucleus merely sequestered the unused morphogenetic determinants (Sander, 1986; Gilbert, 1988). Several geneticists, however, claimed that differentiation and morphogenesis were just epiphenomena of gene activity and that development was equivalent to differential gene expression. In any synthesis of molecular genetics and embryology, the new trinity of DNA, RNA, and Protein has to be reconciled with the older trinity of Ectoderm, Mesoderm, and Endoderm. This is the conceptual problem that we will be the focus of this paper.

The second dimension to any synthesis is the professional dimension. Genetics and embryology diverged in the 1920s and 1930s. The mechanism of this separation has been addressed elsewhere (Gilbert, 1978; 1988; Sander, 1986; Allen, 1986) and will be mentioned but briefly here, for this was not an amicable separation. At the turn of the last century, genetics and embryology had been joined in a common science of heredity. The speculations of August Weismann, W. K. Brooks, Theodore Boveri and others each assumed that the hereditary and developmental determinants were the same. What was inherited was a mode of development. Even Mendel's term for his factors, Formbildungelementen, displays this union. T. H. Morgan, an embryologist, had initiated experiments to show that the cytoplasm, not the nucleus, determined the sexual phenotype of the organism. His results, reported first in 1911, however, showed that the chromosomes were the important determinants not only of sex, but of other traits as well. Although originally presented in embryological terms, Morgan's laboratory gradually refined the gene concept, and by 1926, Morgan formally separated genetics—the transmission of nuclear genes—from embryology, the expression of those genes (Morgan, 1926a,b).

By the 1930s, genetics and embryology had their own rules of evidence, their own paradigmatic experiments, their own favored organisms, their own professors, their own journals, and most importantly, their own vocabulary. Since then, genetics has become explained in terms of molecular biology, and embryology has become developmental biology. Any resynthesis that attempts to explain development in terms of molecular biology has to overcome entrenched disciplinary boundaries. The professional dimension depends upon the view of the relationship between genetics and embryology. If genes only put the final touches on an organism's phenotype, genetics would just be part of commercial animal and plant breeding.

Third, there is the philosophical dimension to the synthesis of embryology and molecular biology. They are as far apart on the philosophical spectrum as they can be. The geneticists of the 1930s and the molecular biologists of the 1980s tend to be reductionists. Historians (Roll-Hansen, 1978; Allen, 1985) and philosophers (Wimsatt, 1984; Darden, L., 1991) have documented the reductionist philosophy of the Morgan school of genetics, and the same philosophy is seen in their molecular descendents. If you know the rules for the "genetic program," the development and evolution of the organism can be understood. In this way,"genetic program" becomes synonymous with mechanism of development. Embryologists, on the other hand, have a reputation for being holistic thinkers (Harwood, 1993; Gilbert and Faber, in press). Embryologists have historically stressed that "the organism in its totality is as essential to an explanation of its elements as an explanation of its elements are to an explanation of the organism" (Haraway, 1976). Embryologist Paul Weiss (1968) thought it philosophically untenable that the organization of the embryo could arise without pre-existing order "The true test of a reductionist system," he said, "is whether or not an ordered unitary system...can, after decomposition into a disorderly pile of constituent parts, resurrect itself from the shambles by virtue solely of the properties inherent in the isolated pieces." He graphically illustrated this point (Weiss, 1962) by showing a photograph of an intact chick embryo, a chick embryo that had been blended through a homogenizer, and a chick embryo whose homogenized components had been centrifuged. The problem for reductionists, he maintained, was how to get that chicken back.

Pnina Abir-Am (1991) has shown the difficulties that Joseph Needham experienced in his attempt to try to reconcile the holistic embryology of the 1930s with reductionist chemistry. The chemists saw the constantly changing embryo as poor material from which they could isolate and characterize their enzymes. Embryologists saw Needham's biochemistry as a reductionist threat. Any synthesis of molecular biology and embryology must be able to reconcile reductionism with holism or else subjugate one by the other. Biochemistry and molecular biology are both reductionist in that they both attempt to explain living processes solely in terms of chemical and physical principles. They start, however, from different premises. The biochemists have traditionally maintained that metabolism is the sine qua non of life (and that the cell is therefore the simplest organism), whereas molecular biologists have seen replication as the fundamental property of life (and elect the virus as life's simplest form). See Gilbert, 1982. Embryology retained a wholism that was in fundamental disagreement to this approach (see Gilbert and Faber, in press).

Another philosophical difference concerns epigenesis and preformationism. E. B. Wilson said in 1925 that "Heredity is effected by the transmission of a nuclear preformation which in the course of development finds expression in a process of cytoplasmic epigenesis." Note that he was using the old termoheredityowhich incorporates both genetics (which stresses the transmission of preformed genes) and embryology (which stresses the epigenetic changes that create new cell types and organs from the mitotic descendents of the fertilized egg). How can these identical preformed genes create such divergent cell types? Several embryologists doubted they could. In 1941, N. J. Berrill, presiding over the first Growth Society meeting (which was to become the Society for Developmental Biology), defined genes as "statistically significant little devils collectively equivalent to one entelechy." Similarly, the contemporary developmental biologist Lauri SaxEn (1973) has claimed that "Our present idea of progressive differentiation actually is not far removed from this classical homunculus concept. Thus, all the information required to build a complete organism is already present within the zygote and development is seen as a progressive expression of this genomic information." He satirized this view by comparing a "'homunculus' in the sperm as illustrated by the 16th cenury animalculists" with "the present view of the 'homunculoid' information in a germ cell." Susan Oyama (1985) and embryologist H. F. Nijhout (1990) have also commented extensively on the similarity in the modern use of "genetic information" and "genetic program" with older concepts of entelechy and preformation.

In recent years, there has been more criticism of the genetic program model. Embryologists are now experimenting with new versions of the morphogenetic field concept, and the notion that these fields are the fundamental unit of development and evolution is being used by a former student of Berrill, Brian Goodwin (1982), Eddy De Robertis and colleagues (1991), and John Opitz (1993). Indeed, one cannot clone or make an antibody to the genetic program. No such thing exists. The genome is less like a programmed score than it is an orchestra wherein each member plays a single note and has perfect hearing. Upon hearing a certain phrase, a performer plays its note, which becomes part of a new phrase, et cetera.

Fourth, there is an aesthetic dimension. Embryology has a tradition of celebrating the complexity and diversity of life. Molecular biology has a tradition of celebrating life's underlying unity and simplicity. Embryologist Berrill, for instance (1961) writes repeatedly of "the amazing diversity of developmental performances" and "the complex reality" of embryonic development. Molecular biologists such as Monod claim (quoted in Jacob, 1988) that the elephant is constructed on the same principles and using the same materials as E. coli. The aesthetic of molecular biology is abstract formalism. Like abstractionist art, molecular biology seeks to get past the apparent diversity of nature to reveal an underlying unity "more real than the real." Just as an abstract painter might represent a table by a line without concern as to whether the table is oak, plastic, metal, red, or white, so molecular biology has traditionally ignored species differences to discover the underlying unities of living organisms. The traditional asthetic of embryology has been naturalism. Every species develops in a different manner, and generalizations from one species to another are very risky. So aesthetically, embryology is to molecular biology as a Michaelangelo statue is to a Brancussi. Needham (1932) was probably correct when he depicted biology as contested ground between Aristotle and Plato.

II. Embryology and the genetic redefining of development

One of the reasons that molecular biology has taken so long to enter embryology is the longstanding fear among embryologists that genetics—whether it be classical genetics or molecular biology—is trying to take over their discipline and bring with it all its reductionism and lack of appreciation for the complexity and species differences. The fear that developmental biology might be "taken over" by genetics is nearly as old as the separation of genetics from embryology in the 1920s.

The remarkable success of genetics in the 1920s and 1930s caused it to become the preeminent way to study inheritance, and it redefined the other disciplines in genetic terms. The study of inheritance became genetics, which Morgan defined as the discipline concerned with the transmission of nuclear genes (Morgan, 1926a,b), and Morgan's exclusion of cytoplasm from the realm of inheritance was soon viewed as dogma (Sapp, 1987). Embryology was redefined as the study of changes in gene expression over time (Morgan, 1934), and evolution was redefined as changes in gene frequency over time (Dobzhansky, 1937). Thus, evolution and embryology, which had traditionally been sciences of the phenotype, were given new, genotypic, definitions.

These new definitions went against the prevailing paradigms of these fields. Evolution had been the province of paleontologists who reconstructed ancient skeletons and phylogenies. Similarly, few embryologists had concerned themselves with questions of gene expression. The predominant problem of embryology from the 1700s through the 1950s was the creation of ordered form, morphogenesis, not differentiation (Haraway, 1976; Lenoir, 1982; Fischer and Smith, 1984). Morphogenesis was a whole-embryo question, differentiation a cellular question. The genetic redefinition of embryology collapsed the morphogenesis question into a subset of the cell differentiation question. To geneticist Richard Goldschmidt, this was axiomatic. He wrote in 1939, "Development is, of course, the orderly production of pattern, and therefore after all, genes control pattern." Similarly, Sturtevant (1932) told the International Congress of Genetics:

One of the central problems of biology is that of differentiation—how does an egg develop into a complex organism? That is, of course, the traditional major problem of embryology; but it also appears in genetics in the form of the question, "How do genes produce their effects?"

Note how embryology's main question is said to be differentiation, not morphogenesis. Note also how genetics can now take over this question. Embryology is being reduced to genetics. And why not, since the geneticists will write that they have the superior methodology. After this separation of genetics from embryology, the geneticists defined themselves against embryology. They often depicted embryology as being an intellectual backwater and saw themselves as the true heirs of Darwin and the question of heredity. William Bateson, who was fond of Biblical rhetoric, cast the split in terms of faith and truth. Embryology had lost the faith, while genetics carried it forward. In his essay of 1922, "Evolutionary faith and modern doubts," he claimed, "Morphology having been explored in its minutest corners, we turned elsewhere...The geneticist is the successor of the morphologist." As the victorious geneticists wrote the history, this historiography became the received doctrine. We have been told that genetics grew as embryology withered on the vine. Even embryologists started believing it. There is a contemporary developmental biology book that presents the history of developmental biology in terms of "Embryology: Losing the Faith" and "Genetics: Keeping the Faith."

The relationship between molecular biology, molecular genetics, and genetics is complex. I agree with Richard Burian's claim that molecular biology is a set of techniques and assumptions that can span the disciplines. As genetics became more and more restricted into the gene mapping program, Morgan portrayed embryology as more and more muddled. Morgan's 1932 essay in Science became the centerpiece for the history of American genetics, and its supersessionist view of genetics rising above embryology has been perpetuated until very recently.

But while the geneticists were making their great discoveries into the mechanism of hereditary transmission, the embryologists were also having their own golden era. Ignoring genetics altogether, embryologists embarked on the program which Joseph Needham (1936) christened "Gestaltungsgesetze, the rules of morphological order." Here, the transplantation experiments of Spemann, the Mangolds, Holtfreter, Hamburger, Hörstadius, Harrison, Witschi, Leewis, Child, Willier, and Rawles set new experimental standards for embryologists and provided astounding new insights into how organs were constructed. The evidence gave coalesced into a concept called the morphogenetic field, and this morphological unit, rather than the gene, was seen as being the fundamental unit of development (and in the case of planaria, inheritance) (Huxley and De Beer, 1934; Weiss, 1939). Indeed, the concept of gene expression is absent in the major embryology books of the 1920s through the 1940s (see Gilbert, 1988). Although experimental embryology had successfully separated itself from the earlier traditions of developmental anatomy, it remained a phenotypic science, and it identified itself as a science concerned with cytoplasmic changes. As Frank R. Lillie wrote in his critical review of 1927, "The germ exhibits the duality of nucleus and cytoplasm; the geneticist has taken the former for his field, the embryologist the latter."

However, the nuclear envelope proved to be a permeable barrier. More and more, geneticists began to see some form of differential gene activity as the cause for embryogenesis. Jumping over the nuclear boundary, they claimed embryology as part of their domain as well. Morgan's group (Dobzhansky, Sturtevant, Schultz, and the visiting Waddington) and the German groups (Kuhn's laboratory, Goldschmidt's laboratory) began studying the mutations which altered the basic patterns of insect development (Harwood, 1993; Kohler, 1994). Goldschmidt (1939, p. 1) saw the development as being identical with "physiological genetics." He claimed that geneticists must explain embryology because the embryologists were not capable of doing so. In a later statement that reflects this boundary dispute, Goldschmidt wrote (1955) that "geneticists will continue to worry about the problem of genetic action and take the risk of climbing over the fence erected by some jealous embryologists, who, while claiming the kingdom for themselves, do not set out to till its soil." C. H. Waddington (1939) began reintroducing embryology into English-language genetics textbooks by stating, "Now that the mechanism of inheritance is known, in its main outlines at least, it is possible to tackle the next question, of how the genes affect the developmental processes which connect the fertilized egg into the adult organism." If the embryologists were not going to discuss embryogenesis in terms of gene activity, the geneticists would.

But the embryologists had a strong research program of their own, and they did not like being told how to do their science. Ross Harrison (1935), Chairman of the section of zoological sciences of the American Association for the Advancement of Science, addressed his colleagues in words evoking the political anxieties of the mid-1930s:

Now that the necessity of relating the data of genetics to embryology is generally recognized and the "Wanderlust" of geneticists is beginning to urge them in our direction, it may not be inappropriate to point out a danger of this threatened invasion. The prestige enjoyed by the gene theory might easily become a hindrance to the understanding of development by directing our attention solely to the genome, whereas cell movements, differentiation, and in fact all of developmental processes are actually effected by the cytoplasm.

He despaired that identical genes could produce the different types of cells seen in the adult organism.

III. Redefining differentiation: The adaptive enzymology of development

Genes could not be relevant to embryology until geneticists could explain how identical genes could generate different types of cells. In 1934, T. H. Morgan speculated that every gene, even though present in every cell, might not be active in every cell. Forty years earlier, as a newly minted embryologist, Morgan had worked with Hans Driesch, the leading proponent of epigenesis in his time. Driesch (1894) had just finished writing his Analytische Theorie de organischen Entwicklung. Here, he put forth the following view for the mechanism of epigenesis:

Insofar as it contains a nucleus, every cell, during ontogenesis, carries the totality of all primordia; insofar as it contains a specific cytoplasmic cell body, it is specifically enabled by this to respond to specific effects only. ... When nuclear material is activated, then, under its guidance, the cytoplasm of its cell that had first influenced the nucleus is in turn changed, and thus the basis is established for a new elementary process, which itself is not only the result but also a cause.

In 1934, Morgan dusted off his copy of Driesch's book and refers to it when he updates the account of epigenesis:

The initial differences in the protoplasmic regions may be supposed to affect the activity of the genes. The genes will then in turn affect the protoplasm, which will start a new series of reciprocal reactions. In this way we can picture to ourselves the gradual elaboration and differentiation of the various regions of the embryo.

In other words, cells that contain different types of cytoplasm would be able to activate different batteries of genes. Here we also see, in its anlagen, the notion of feedback between nuclear genes and cytoplasmic proteins which will become the hallmark of the operon-models of eukaryotic gene expression. Other biologists, notably Haldane (1932) and Goldschmidt (1939) were also postulating that genes acted at different times during development; but they did not place so much value on the cytoplasmic feedback.

This may be a pleasant hypothesis, but there was no evidence in its favor. Indeed, there were other models as consistent with the data. Richard Goldschmidt (1939) proposed that timing differences in the genes determined the phenotype of the cell. Ernest Everett Just (1939) proposed that the cytoplasm contained all the developmental determinants and that the nucleus was like a refuse bin for unused determinants. In the late 1930s, even Curt Stern admitted that this idea could not be disproven (see Gilbert, 1988).

The evidence that differentiation is caused by the differential expression of batteries of genes from identical genomes came from the study of unicellular organisms. As Burian pointed out, Hämmerling's experiments on Acetabularia checked some of these models and showed that the nucleus produced substances that were essential in constructing the cellular phenotype, at least in unicellular organisms. But an explanation of how identical genes could yield different cell types came from studies of enzyme synthesis in yeast and E. coli.

I would like, therefore, to pick up this story in 1947; for here is where the discourse on development changes from the tissue level to the intracellular level. Here is where differentiation becomes defined as changes in cytoplasmic proteins, and adaptive enzyme synthesis becomes used to model metazoan embryogenesis. The explicit linkage of enzymatic adaptation and cellular differentiation was made by Jacques Monod at the Growth Society meeting of that year. Monod introduced the phenomenon of enzymatic adaptation as a possible solution to the problem of how identical genomes can synthesize different "specific" molecules:

The widest gap, still to be filled, between two fields of research in biology, is probably the one between genetics and embryology. It is the repeatedly stated—and thus far unsolved problem—of understanding how cells with identical genomes may become differentiated, that of acquiring the property of manufacturing molecules with new or, at least, different specific patterns or configurations.

Monod, however, did not feel that there was enough data from embryonic cells to warrant concluding that the same mechanisms guided both microorganisms and eukaryotic embryos.

A more systematic mapping of the phenomena of enzymatic adaptation to differentiation was made at the symposium of the Society for Experimental Biology that year, "Growth in Relation to Morphogenesis" (Danielli and Brown, 1948). Numerous topics were discussed. There were several papers on the roles that hormones have in regulating cell growth and development in animals and plants. Two other papers, those by Johannes Holtfreter on neural induction in amphibians and by H?rstadius and Gustafson on sea urchins, discussed the biochemistry of animal development. Hans Grneberg and Ernst Hadorn spoke on mutations that effected the embryos of mice and Drosophila, respectively; and three papers dealt specifically with the question of how the nuclear genotype could produce different types of cells. These three papers reached a remarkable consensus and represented the approaches of a microbial geneticist, a vertebrate embryologist, and a plant physiologist.

The most critical of the papers was that of Sol Spiegelman of the Washington University School of Medicine. Expanding on a paper that had been published the preceeding year, Spiegelman (1948) starts his paper by redefining the problem of differentiation in biochemical terms. First, the problem of differentiation must move from being seen as a morphological property of tissues to being seen as a biochemical property of individual cells. Differentiation is to be seen not in terms of tissue structure but "as the controlled production of unique enzyme patterns." This redefinition, he stated, would focus our attention on "the relationship between the genes in the nucleus and the properties of the cytoplasm."

One of the two starting points for Spiegelman's synthesis is the Beadle and Tatum's studies in Neurospora that showed that an altered genotype creates altered enzymatic properties in the cytoplasm. Enzymes were the link between genotype and phenotype. Genes controlled the production of enzymes; enzymes controlled the phenotype of the cell. The second source of Speigelman's theory is the work done by J. P. Greenstein and colleagues that showed that the cells of each tissue had different patterns of enzymes. Indeed, this is what would be expected if differentiation were the regulation of unique constellations of enzymes.

The next question, then, was how could the supposedly identical genome of each cell type be regulated to produce different constellations of enzymes in different types of cells. The answer, claimed Spiegelman, might come from the study of adaptive enzymes in microbes such as yeast. Here are instances where the identical genome produces different enzymes under different environmental conditions. "A population of individual cells placed in contact with a particular substrate acquires, after some lapse of time, the enzymes needed to metabolize the substrate". Moreover, "once the enzyme has been induced, its maintainance requires the continued presence of the substrate." The use of the term induced here is critical, for it links this effect directly to embryology. While the phenomenon of adaptive enzymes had been known since the turn of the century in a variety of microbes, Spiegelman claimed that the yeast system is the best since the cells can be cultured in nitrogen deficient medium which would prevent their growth. When the enzymes are assayed, one could be certain that the genome has not changed nor were rare variants selected by differential growth on the substrate.

So Spiegelman used the production of adaptive enzymes by yeast cells as a model for metazoan embryos. One could certainly study the biochemistry of cells much better in the clonal microorganism than in the constantly changing embryo! What did Spiegelman find? First, he showed that the inducer could cause identical genomes to synthesize different proteins. Second, he showed that the kinetics of enzyme production approximated the curve expected of an autocatalytic reaction. The amount of β-galactosidase in yeast after induction with lactose was almost exactly what was predicted by equations in which the velocity of enzyme synthesis was a function of the amount of enzyme already present (i.e., an autocatalytic synthesis).

New enzyme production was not merely a consequence of activating an inactive precursor, stabilizing an unstable protein, or creating more enzymes from the nuclear genes. Rather, it appeared that the more enzyme was produced, the faster the reaction proceeded. "On these grounds one is led to propose that enzyme formation is mediated by a mechanism which is inherently autosynthetic...Accepting this assumption of autosynthesis posseses definite implications for the problem of gene action."

In addition to the ideas that differentiation was caused by the controlled synthesis of unique enzyme patterns, and that enzyme synthesis was autosynthetic, Spiegelman added a new notion: competition for scarce resourses. The different enzymatic reactions compete for a limited amount of amino acids and energy. As the synthesis of new adaptive enzyme proceeded, the synthesis of certain other enzyme systems declined drastically. The two alternative biochemical pathways could be in competition with each other. In this model, the "protein molecule was stable, but the enzyme-forming system involved in its formation was a poor competitor for protein material in the absence of substrate. This results in the loss of protein to other synthesizing systems." Spiegelman used the "analogy of predator-prey relationships encountered in the ecology of higher organisms" to make his point. After showing that a trivial explanation, that of enzyme stability being dependent upon substrate concentration, can be excluded, Spiegelman concluded, "Apparently the extent, severity, and kind of competition do, however, determine the types and amounts of enzymes found in the cytoplasm of cells. Consequently, any theories of gene action...must provide a mechanism whereby these agents can perform their functions by influencing the outcome of the competitive interactions amongst enzyme-forming systems. "

This led him into a plasmagene model of differentiation. Using radioactive phosphate as a tracer, Spiegelman showed the correlation that the synthesis of new protein parallels a transfer of radioactive phosphate out of the nucleus and into the cytoplasm and that this radioactive phosphate binds to proteins other than that being synthesized. These correlations had also been made that year independently by Caspersson and by Brachet. Based on the new research by Lippmann and by the Cori's, Spiegelmann postulated that the nucleoproteins were energy donors that make specific protein synthesis possible. (He notes that H. J. Muller had independently suggested the same conclusion). These nucleoprotein units were the plasmagenes, and they were the competitors in the ecological drama of differentiation. The genes of the nucleus produced the plasmagenes which migrated into the cytoplasm. Like the genes, these plasmagenes were also self-duplicating entities. Moreover, they were able to synthesize enzymes. Substrates were envisioned to stabilize the plasmagene complex enabling them to duplicate and to make more enzymes. Here, the plasmagene-enzyme complex is labile unless stabilized by the substrate. This hypothesis "would suggest that nucleoprotein is the most active fraction of the cell in stimulating and guiding differentiation." Moreover, "the fate of any given cell during morphogenesis will be determined by the outcome of the competitive interactions amongst its original plasmagenic population. In these terms, one can understand the multipotency of the early embryonic cell as well as the restriction of morphogenic plasticity which accompanies the progress of differentiation."

Spiegelman concluded that one could readily rewrite experimental embryology texts in terms of plasmagene theory, but that it would do little good, since so very little is known. Nevertheless, Spiegelman is confident that expressions such as competence, induction, evocation, and biological fields could be given more precise meanings in this new molecular scheme.

And it seems that at this point, embryologist C. H. Waddington concurred. Waddington had been active in attempting to synthesize genetics and embryology for the past decade. He published one of the first papers linking genes to phenotypic defects observable in embryos, and he championed the concept of competence whereby multiple pathways of reactions would exist in a tissue (Waddington, 1940a; 1940b; see Gilbert, 1992). The inducer merely caused the cell to follow one of these pathways and not others. He had already linked genetics to embryology by showing that certain mutations could cause alterations in cell fate analogous to those caused by inducer tissues.

Waddington (1948) maintained that all the events of morphogenesis were gene-directed processes. Even tissue formation that may result from differential cell stickiness, he claims, depends upon genes that make the substances that cause the cells to become adhesive. However, while Spiegelman stressed different protein constellations as the end result of differentiation, Waddington saw cytoplasmic differences as also being the cause of differentiation. This view of a reactive genome that could be influenced by the cytoplasm characterized certain embryologists of the 1940s and 1950s, especially he and Paul Weiss. As we will see, the ability to see genes as a library of potencies acted upon by cytoplasmic factors was important in Waddington's later adoption of the operon model.

Differentiation, in Waddington's view, was self-reinforcing, irreversible, and canalyzed. By the last term, he meant that development could only proceed down certain paths. Not all potencies were realized. A cell might become a nerve or a skin cell depending upon whether it is induced by the notochord. It does not become anything that is both. (When it does, it may be malignant). Given these properties of development, how might differential cytoplasmic synthesis be established and maintained?

Waddington's solution, for which he gives nobody else credit and which he presented as his own, is a competition model for the resources of the cell. In a cell which has not yet been epigenetically determined as either neural or epidermal, the gene complement must endow it with the potencies for carrying out the synthesis of either of the two types of proteins; there must, in fact, be alternative chains of synthesis, each leading from certain genes to certain cytoplasmic proteins, which are in competition for much the same substrate materials. In such circumstances, comapartively slight changes in the available raw materials might shift the whole dynamic system from one path into another. The progressive substrate changes would lead to changes in new enzyme synthesis, which would, in turn, create further substrate changes.

But "something further is necessary to account for the self-reinforcing character of differentiation." Waddington reminds us that before overt differentiation, there is a stage called determination where obvious differences are not seen, even though the cells are committed to become one cell type and not another. The cells become progressively different from each other. The simplest explanation of this, he claims, would be if "protein synthesis involves an autocatalytic event." These autocatalytic events, moreover, would be in competetion with one another. "If one has a number of autocatalytic synthetic processes competing for the same collection of substrates...it seems inevitable that there would be certain sets of syntheses which were compatable in the sense that they could go on simultaneously, while other sets would be incompatable and unrealizable."

This would lead to the canalization of development and would explain the stability and irreversibility of differentiation. This is what Spiegelman had also noted. Waddington noted that others such as Darlington, Lindegren, Spiegelman, and Sonneborn, have proposed such self-duplicating plasmagenes on genetic evidence, but that this type of entity fit in well with the observed facts of embryology.

The paper by the botanist, K. Mathers (1948), gave a different perspective on the plasmagene theory. His plasmagenes were critical for the continuity of cytoplasmic function in the absence of the genome. Like Waddington (who uses snail coiling mutants to the same end), Mathers found that there were cases like that of pollen sterility in Nicotinia where the effect was seen in the absence of the nuclear genes controlling this effect. However, upon analysis, the cytoplasm was seen to be given these traits by the parental nucleus. Similarly, Mathers quoted Hämmerling's experiments on Acetabularia to the effect that the nucleus controled the type of cap that is formed in this alga, but the cytpolasm was a reservoir of morphogenic substances. There were plasmagenes, to be sure; and they were seen as gene products that reproduced themselves in the cytoplasm under certain chemical conditions. Although these plasmagenes were seen as mediators between the genome and the cell phenotype. But since plant development is not characterized by the irreversibility and stability of differentiation seen in animal development, the competitive interactions postulated by Spiegelman and Waddington were downplayed. The chemical conditions that would enable one plasmagene to multiply whereas another would not are not discussed. So having stated the need for plasmagenes, he did not consider them as playing major roles in cell determination.

Mather saw the the plants' developmental canalization as being influenced more on the nuclear level. Unlike Waddington and Spiegelman, Mather believed that the genome is fluid and that the association of a gene with other genes could influence its activity. Moreover, unlike the other two, Mather thought that whether one gene produced one or many products was still an open question. Different amounts of the cytoplasmic raw materials could influence the nature of these products and the arrangement of the gene with other genes. He postulated, like Waddington did, that the genes both reacted to the cytoplasmic molecules in their environment and created new cytoplasmic environments by this action. All three papers saw adaptive enzymes as a fitting model for cell differentiation in multicellular eukaryotes.

IV. Enzyme induction and Embryonic Induction

The phenomena of adaptive enzymes was brought closer to embryology by a nomenclature agreement published in Nature, 1953. Here, the major researchers in the field, M. Cohn and J. Monod of the Institut Pasteur, M. R. Pollock of the National Institute for Medical Research in London, S. Spiegelman, now at the University of Illinois, and R. Y. Stanier of the University of California at Berkeley agreed to a uniform terminology for enzyme formation (Cohn et al., 1953). No longer would the term adaptive enzyme be used, for the term "adaptive" denoted a change which increased the evolutionary fitness of the organism. Rather, the process would be known as enzyme induction and "any substance thus inducing enzyme synthesis is an enzyme 'inducer.' An enzyme-forming system which can be so activated by an exogenous inducer is 'inducible'." At the end of the letter, the signatories note that "the exposure of an organism to a single inducer which is also a substrate may result in the induction of a sequence of enzymes, since the metabolism of the primary inducer gives rise to the formation of a succession of intermediary metabolites each of which serves as an inducer for the enzyme which converts it to the next member of the metabolic chain. This phenomenon is termed 'sequential induction' (simultaneous or successive adaptation)."

This letter takes the microbial enzyme formation away from evolutionary and ecological biology and closer to embryology. Induction was, of course, the key concept of vertebrate embryology, and successive (secondary) inductive chains were the mainstay of embryology texts. The notochord induced the neural plate which induced the lens which induced the cornea, etc. Of chief concern was the "primary" inducer, the Organizer, discovered by Spemann and Mangold in 1924. According to Lvtrup and colleagues (1978), "few compounds, other than the philosopher's stone, have been searched for more intensely than the presumed agent of primary embryonic induction in the amphibian embryo," and Harrison (quoted by Twitty, 1966) referred to the amphibian gastrula as a "new Yukon to which eager miners were now rushing to dig for gold around the blastopore." The concept of induction was at the core of both the morphogenesis and differentiation question and was probably the most powerful principle in embryology during the 1920s through 1950s. The linkage of microbial enzyme induction to this central concept of embryology was an extremely important devise in asserting the relevance of microbiology for unravelling the problems of development (and perhaps cancer as well), and for asserting the fundamental unity of all living things. The relevance of adaptive/inducible enzymes to cancer research has been a point that microbial biologists strove to mention. At times, the connection has been very week. In 1946, Spiegelman and Kamen will point out, "The problem of cancer involves explaining the appearance of a sudden hereditable change in somatic cells analogous in several ways to enzyme adaptation or cellular differentiation." These ways were not listed. The success of this approach is appreciated by this author who was for two years privileged to study E. coli ribosome synthesis on an NIH cancer grant.

Although the concept of adaptive enzymes had been linked to embryological induction, there had been no great technical or conceptual advances that were immediately made. Curt Stern (1955) reviewed the data on gene synthesis of proteins and found the situation very similar to what it had been in 1947. He agreed with Mather that the identity of the genome in differentiated cells was an unproven speculation and that slight modifications of the genome could indeed have occured during development. He summarizes the competition model, but leaves out the plasmagenes. The cytoplasm is seen to act directly on the genes and noton their intermediate products. After reviewing enzymatic adaptation (and calling it that), Stern claims that there is no convincing evidence for the existence of the postulated plasmagenes.

However, by 1958, the idea that microbial enzyme synthesis could model eukaryotic development had been digested, processed, and had elicited a reaction. In this year, a symposium on the Chemical Basis of Development was conducted at the Johns Hopkins University. It was an eclectic gathering, and there was much discussion of induction, both embryonic and microbial. By this time, another line of evidence has emerged against the plasmagene/enzymatic adaptation view of differentiation. The inducers of the microorganismal enzymes are small substrate molecules. But evidence from Toivonen, Kuusi, Yamada, and Niu was suggesting that the inducer of the amphibian neural tissue was a protein. Clement Markert concluded that there was very little evidence in embryos for the type of adaptive enzyme synthesis found in microbes (1958; p. 6). Nor was there any particular eviudence in favor of plasmagenes. Markert thought that differentiation could be measured by the constellation of proteins in the cell, but his hypothesis of differential gene activation is pretty much the same as Morgan framed it in 1934. However, he is able to buttress the notion of genetic identity of each nucleus with the results of King and Briggs on frog nuclei and Pavan and Beerman on Drosophila chromosomes. Markert's model of differentiation concerns cytoplasmic proteins that are able to enter the nuclei and bind to the chromosome in order to activate specific genes.

But there are plenty of speakers who supported the idea that embryonic induction can be modelled by microbial induction, and Bentley Glass (1958) noted that "inasmuch as some of the same enzyme systems significant in vertebrate development can be readily followed in bacteria, the approach holds considerable appeal." Glass criticised Markert for not recognizing that suppressor genes can shift the bacterial phenotype from one state to another. Melvin Cohn (1958) discussed the b-galactosidase inducible system of E. coli, and discussed it in terms of embryonic development. "If maintainance in bacteria is to enjoy the dignity of being compared to differentiation in higher organisms, then it should persist over many generations; and in fact it does." Other analogies to metazoan development are also made, such as heterogeneity of response. The ability to respond to inducer was found to be dependent upon the permease that enables the inducer to enter the cell. This is related to embryonic competence, although not named as such.

Three other papers, by L. Gorini and W. K. Maas, by H. J. Vogel, and by Boris Magasanik, respectively, looked at feedback mechanisms of regulation in bacteria. Vogel (1958) speculated that the inhibition of the synthesis of some product might actually be due to the product's binding to the gene and shutting it off. Here, then, was another example of differential protein synthesis whereby a small molecule (in this case, the product of a series of metabolic reactions) interacted with the genes to alter their activity. Gorini and Maas (1958) speculated that a "change in the enzymatic constitution resulting from either feedback inhibition or enzyme induction could be the step initiating the series of reactions necessary to produce a differentiated cell." Boris Magasanik (1958) concurred, saying that it was "not unlikely" that cellular differentiation could be controlled in a manner similar to that of negative feedback regulation.

But does anything like this ever happen in real embryonic cells? A paper by Richard Stearns and Adele Kostellow of the Albert Einstein College of Medicine (1958) provided evidence that enzyme induction in E. coli and enzyme induction in embryonic blastomeres could be one and the same. They purposefully attempted to mimic adaptive enzyme synthesis in embryonic cells. First, they made Rana pipiens embryos more like bacteria and yeasts by dissociating them into blastomeres with versene (EDTA). They then centrifuged them according to density to get mixtures enriched for surface cells, presumptive ectoderm, presumptive mesoderm, and presumptive endoderm. In normal embryos, tryptophane peroxidase is found only in the liver, an endodermal organ. Moreover, the levels of the enzyme could be induced in the liver. When they determined if tryptophane peroxidase could be induced by tryptophane in embryonic cells, they find that no cell could be induced before gastrulation, but after gastrulation, the gut precursor cells show induction. Moreover, lactose was found to be able to induce b-galactosidase. It seemed that, despite Markert's objections, embryonic and microbial induction were the same after all.

It is here, too, that we get to hear one of those statements of impatience raised by molecular biologists when they confront embryologists. In the discussion following Dr. Magasanik's talk, Sol Spiegelman (1958) gave a little oration:

I have found it difficult to avoid the conclusion that many of the investigators concerned with morphogenesis are secretly convinced that the problem is insoluble. I get the feeling that many of the intricate phenomena described are greeted with a sort of glee as if to say, " My God, this is wonderful, it is so complicated we will never understand it."It seems to me that perhaps the time has come to abandon this joyful pessimism and its attendant conviction of incomprehensible complexity. In particular, I should like to make a plea for a more optimistic view based on a belief in simplicity. The phenomena of morphogenesis can hardly be as complicated as implied by the welter of apparently unrelated observations constituting the literature of embryology. ...It is no longer relevant these days to phrase questions of cell physiology in terms of other than chemically defined entities. It seems to me that the same is true for morphogenetic events.

Here we see the Kulturkampf of molecular biology and embryology.

Also by this time, RNA has been discovered and has been found in animal cells as well as in plants. At this symposium, Mirsky and Allfrey reported that while large quantities of RNA are made in the nucleus, only a small fraction is getting into the cytoplasm where the proteins are made. Joe Gall also reported that the visualization by autoradiography of RNA on the lampbrush chromosomes of larval flies and amphibian oocytes. Gall postulated that the locations of the 32P-labeled RNA are genetic loci. By following the tracer, Gall finds that the RNA goes from the nucleus into the cytoplasm.

In the next three years, evidence for the existence of a messenger RNA would increase. Again, this data would come from E. coli and yeast. This RNA, representing less than 5% of the total RNA of the cell would have a short half-life, a rapid incorporation of phosphorus, and the ability to associate with 70S ribosomal particles. Its base composition resembled that of the organism (or in phage-infected bacteria, that of the virus), making it readily distinguishable from the long-lived ribosomal RNAs and from the short soluble RNAs. The ribosome would be established as the site of protein synthesis. The presence of a reusable messenger RNA made the plasmagene hypothesis untenable (Judson, 1979).

These data were synthesized with the ongoing program of Monod's laboratory looking at mutations of the lactose-synthesizing genes of E. coli. The story of the operon model has been told many times before (see Schaffner, 1974; Judson, 1979), and I will not repeat it here except to say that it demonstrated that induction worked at the genomic level. In inductive systems, the inducer blocked a gene-encoded repressor (either RNA or protein) from binding at operator site adjacent to the structural genes. This prevented their activation. If the inducer was present, the gene made its mRNA which bound to the ribosomes to make the proteins. If the inducer was not present, the repressor was able to bind to the gene region and block transcription. In this way, the same genome could give different enzymes depending on whether or not the inducer was present. In their closing statement of a major 1961 review article, Jacob and Monod emphasize that operon-like control mechanisms may be a universal part of gene regulation.

It has been repeatedly pointed out that enzymatic adaptation, as studied in micro-organisms, offers a valuable model for the interpretation of biochemical co-ordination within tissues and between organs of higher organisms. The demonstration that adaptive effects in micro-organisms are primarily negative (repressive), that they are controlled by functionally specialized genes and operate at the genetic level, would seem greatly to widen the possibilities of interpretation. The fundamental problem of chemical embryology is to understand why tissue cells do not express, all the time, all the potencies inherent in their genome.

This fit in perfectly with what Markert had said in 1958 about cytoplasmic proteins binding to regions of the DNA to activate or inactivate specific genes. It also was congruent to the embryologists' view of a reactive genome that would not only produce new cytoplasmic substances but that would take orders from the cytoplasm. In fact, the operon model of development was brought into embryology texts immediately by those people who had been looking for a synthesis of genetics and embryology.

Between 1961 through 1963, at least three major textbooks were published which attempted to synthesize genetics and embryology through molecular biology. Each of them used the operon model, but in different ways. The first book, that of geneticists Ruth Sager and Francis J. Ryan, Cell Heredity (1961), included the information that was to become the operon model. However, as the book had been written before Jacob and Monod's integration, the data were left to stand without being extrapolated to other systems. Rather, they were used to model intracellular enzyme synthesis only. Other eukaryotic regulatory systems (such as those found by McClintock in Maize) were represented as being siimilar. But there was no attempt to model metazoan development by adaptive enzymes.

Waddington's 1962 book, New Patterns in Genetics and Development, however, commenced with a chapter relating the Jacob and Monod operon model to induction neural induction in amphibians. After detailing the genetics of the lac operon, he noted two generalizations that are important for its applicationto embryology. First, there were regulatory genes in addition to structural genes; and second, the gene is the target of the repression or activation, not the protein. Waddington links this to eukaryotic regulatory systems. Like Sager and Ryan, he showed the similarities between the lac operon and the control elements discovered in plants. Next, he quoted the results of Mechelke, who in 1961 correlated the puffing of dipteran chromosomes with ecdysone secretion preceeding pupation. Certain regions of the chromosome puff out before the others, suggesting to Waddington "just the kind of intrachromosomal activity which Jacob and Monod's Operator is supposed to carry out."

Waddington then began to extrapolate from the operon model to metazoan embryos.

If a structural gene controlled by an operator in the first system produced a substance which functioned as a repressor in a second system, we would have the possibility of 'cascade repression'; and if there were a number of links of this kind, complex systems might be built up which exhibit some of the tendency towards irreversibility which is commonly found in embryological systems but which is hardly accounted for on the simple Jacob-Monod scheme.

Waddington is sufficiently aware of history to relate all this back to T. H. Morgan's 1934 statement that the initial protoplasmic regions determine which of the genes are active.

If metazoan embryos utilize operon-like systems for development, two predictions can be made. First, in mosaic embryos, the Jacob-Monod regulators of development should be found in the regionalized cytoplasm of oocytes. Moreover, there should be genes controlling the formation of these repressors. Waddington proposed the deep-orange locus as a candidate for such a gene, even though, he admitted, the data were not totally consistent with this identification.

Second, embryonic inducers might function by blocking a naturally occuring repressor that is already in the competent cell. He claims that he, Needham, and Brachet (and also Holtfreter) were too pessimistic when they found that unnatural molecules could induce neural plate formation. Indeed, they had proposed that the real inducer lay within the competent cell but was bound by an inhibitory molecule. He now modeled induction by two adjacent cells. One cell, the competent cell, could synthesize a molecule that was the functional activator of gene activity. However, this molecule was bound by a repressor formed in the same cell. The inducing cell could synthesize an inducing molecule that acted to remove the repressor from the activator. By modeling embryonic induction by microbial enzyme induction, Waddington was able to harmonize the data showing that the competent cell, itself, had inducing activity and the data that showed that numerous compounds that had no chemical similarity with each other could all cause induction to occur. Some of these molecules would resemble the gene activator, while others would resemble the evocator that separated the repressor from the gene activator.

For Waddington, the operon model vindicated his 1936 hypothesis that the actual inducing molecule was produced by the competent (responding) cell. In his autobiographical notes (Waddington, 1975), he maintained, "We showed that, in these terms, the specificity resides in the cells that react to induction—we called it 'the masked evocator'. This is very similar to the situation discovered by F. Jacob and J. Monod many years later in bacteria, where again the specific repressor molecules are internal to the cells which react to enzyme-inducing substances." As Sahotra Sarhar has pointed out (as editor of this paper), Waddington was able to take the quantitative data of Jacob and Monod's operon model and to distill it to a qualitative model that was applicable to differential gene expression between cells.

In 1963, John Moore published his synthesis of embryology and genetics, Heredity and Development. This book, like Morgan's 1934 volume, was divided into two sections: the first concerned genetics and the second concerned embryology. In the last chapter of the book, he attempted to fuse the two units together. His bridge was the lac operon. "Not only does it satisfactorily account for many genetic facts, but it also provides an obvious way of explaining the role of genes in early development." Embryologists, he wrote, were adamant that non-genetic phenomena could influence what the identical genes actually do, but the geneticists were set against it. "This point of view, which once would have been reasonable to an embryologist but not to a geneticist, now seems reasonable to both." He illustrated this using the example of polarized light and pH (two definitely non-genetic influences) on Fucus development. Indeed, reading Moore, one gets the impression that the lac operon model was a victory of the embryological view over the genetic.

Moore's book consists largely of the geneticists' fly Drosophila and the embryologists' frog Rana pipiens. It ends, however, with a story of unification and harmony effected by the molecular biologists' microbe E. coli.

A generation ago, few embryologists or geneticists would have predicted that a synthesis of their fields would be made possible by studies on the bacterium Escherichia coli. But this microscopic creature, with no embryology of its own, has shown a way. A decade from now it may be difficult to distinguish between a geneticist and an embryologist, as they advance their science beyond what each might independently achieve.

The year 1963 also saw the first experiments using transcription inhibitors to demonstrate the importance of differential gene transcription in the embryo (Nemer et al., 1963; Scott and Bell, 1963). The molecularization of embryology had begun.

Literature Cited

Abir-Am, P. (1991). The philosophical background of Joseph Needham's work in chemical embryology. In A Conceptual History of Modern Embryology (S. F. Gilbert, editor). Plenum Press, New York. Pp. 159-180.

Allen, G. E. (1985). Thomas Hunt Morgan: Materialism and experimentalism in the development of modern genetics. Trends Genet. 3: 151-154, 186-190.

Allen, G. E. (1986). T. H. Morgan and the split between embryology and genetics, 1910-1935. In A History of Embryology (T. J. Horder, J. A. Witkowski, and C. C. Wylie, editors), Cambridge University Press, Cambridge. Pp. 363-395.

Bateson, W. (1922). Evolutionary faith and modern doubts. Science 40: 1412- 1415.

Berrill, N. J. (1941). Spatial and temporal growth patterns in colonial organisms. Growth Symposium 3: 89-111.

Berrill, N. J. (1961). Growth, Development, and Pattern. Freeman, San Francisco.

Browder, L., Erickson, C. and Jeffreys, W., Developmental Biology, third edition. Saunders, Philadelphia.

Cohn, M. (1958). On the differentiation of a population of Escherichia coli with respect to b-galactosidase formation. In A Symposium on the Chemical Basis of Development (W, D. McElroy and B. Glass, editors). Johns Hopkins Press, Baltimore, MD., p. 459.

Cohn, M., Monod, J., Pollock, M. R. , Spiegelman, S., and Stanier, R. Y. (1953). Terminology of enzyme formation. Nature 172: 1096.

Danielli, J. F. and Brown, R. (editors), 1948. Growth in Relation to Differentiation and Morphogenesis. Cambridge University Press, Cambridge.

Darden, L. (1991). Theory Change in Science: Strategies from Mendelian Genetics. Oxford University Press, NY.

De Robertis, E. M., Morita, E. A. and Cho, K. W. Y. (1991). Gradient fields and homeobox genes. Development 112: 669-678.

Dobzhansky, D. (1937). Genetics and the Origin of Species. Columbia University Press, NY.

Driesch, H. (1894). Analytische Theorie der organschen Entwicklung. W. Engelmann, Leipzig.

Fischer, J.-L. and Smith, J. 1984. French embryology and the "mechanics of development" from 1887 to 1910: L. Chabry, Y. Delage and E. Bataillon. Hist. Phil. Life Sci. 6: 25-39.

Gilbert, S. F. (1978). The embryological origins of the gene theory. J. Hist. Biol. 11: 307-351.

Gilbert, S. F. (1982). Intellectual traditions in the life sciences: Molecular biology and biochemistry. Perspect. Biol. Med. 26: 151-162.

Gilbert, S. F. (1988). Cellular politics: Ernest Everett Just, Richard B. Goldschmidt, and the attempts to reconcile embryology and genetics. In The American Development of Biology (R. Rainger, K. B. Benson, and J. Maienschein, editors). University of Pennsylvania Press, Philadelphia, PA.

Gilbert, S. F. (1992). Induction and the Origins of Developmental Genetics. In S. F. Gilbert (editor), A Conceptual History of Modern Embryology. Plenum Press, NY. Pp. 181-206..

Gilbert, S. F. (1992). Cytoplasmic action in development, Quart. Rev. Biol. 66: 309-316.

Gilbert, S. F. and Faber, M. Looking at embryos: The visual and conceptual aesthetics of emerging form. In Aesthetics and Science: The Elusive Synthesis (A. I. Tauber, editor). (Accepted for publication).

Glass, B. (1958) A summary of the McCullom-Pratt Symposium on the chemical basis of development. In A Symposium on the Chemical Basis of Development (W, D. McElroy and B. Glass, editors). Johns Hopkins Press, Baltimore, MD., p. 856-858.

Goldschmidt, R. B. (1939). Physiological Genetics.. McGraw-Hill, NY.

Goldschmidt, R. B. (1955). Theoretical Genetics. University of California Press, Berkely, CA.p. 247.

Goodwin, B. C. (1982). Development and evolution. J. Theoret. Biol. 97: 43-55.

Gorini, L. and W. K. Maas (1958). In Chemical Basis of Development, op. cit., p. 459-478.

Haldane, J. B. S. (1932). Time of action of genes and its bearing on some evolutionary problems. Amer. Nat. 66: 5-24.

Haraway, D. J. (1976). Crystals, Fabrics, and Fields: Metaphors of Organicism in Twentieth Century Developmental Biology. Yale University Press. New Haven.

Harrison, R. G. (1935). Embryology and its relations. Science 66: 361-368.

Harwood, J. (1993). Styles of Scientific Thought: The German Genetics Community 1900-1933.. University of Chicago Press, Chicago.

Huxley, J and De Beer, G. R. 1934. The Elements of Experimental Embryology. Cambridge University Press, Cambridge.

Jacob, F. (1988). The Statue Within. Basic Books, NY. 1988.

Jacob, F. and Monod, J. (1961). Genetic regulatory mechanisms in the synthesis of proteins. J. Mol. Biol. 3: 318-356.

Judson, H. F. (1979). Eighth Day of Creation. Simon and Schuster, NY

Just, E. E. (1939). The Biology of the Cell Surface. Blakiston, Philadelphia.

Kohler, R. 1994. Lords of the Fly. University of Chicago Press, Chicago.

Lenoir, T. (1982). The Strategy of Life. Reidel, Dordrecht.

Lillie, F. R. (1927). The gene and the ontogenetic process. Science 66: 361-368.

Løvtrup, S., Landström, U., and Løvtrup-Rein, H. 1978. Polarities, cell differentiation, and induction in the amphibian embryo. Biol. Rev. 53: 1-42.

Magasanik, B. (1958). In Chemical Basis of Development, op. cit., p. 485-490.

Malacinski, G. M. (1990). Cytoplasmic Organization Systems. McGraw-Hill, NY.

Markert, C. (1958). Chemical concepts of cellular differentiation. In Chemical Basis of Development. Op. cit. p. 6.

Mather, K. (1948). Nucleus and cytoplasm in differentiation. In Danielli and Brown, Growth in Relation to Differentiation and Morphogenesis. Op. cit. pp. 196-216.

Mendel, G. 1865. Versuche über Pflanzen-Hybriden. Verh. naturforsch. Vereines Brünn 4: 1-47 (citation on p. 42). Thanks to J. Opitz for pointing this out to me.

Monod, J. (1947). The phenomenon of enzymatic adaptation and its bearing on problems of genetics and cellular differentiation. Growth Symp. 11: 223-289. I thank Dick Burian for pointing out this reference to me.

Moore, J. A. (1963). Heredity and Development. Oxford University Press. NY. 236.

Morgan, T. H. (1926a). The Theory of the Gene. Yale University Press. New Haven.

Morgan, T. H. (1926b). The genetics and physiology of development. Amer. Nat. 60: 489-515.

Morgan, T. H. (1932). The rise of genetics. Science 76: 261-288.

Morgan, T. H. (1934). Embryology and Genetics. Columbia University Press. NY.

Needham, J. (1932). The Great Amphibian. Student Christian Movement. London.

Needham, J. (1936). Order and Life. Yale University Press. New Haven, CT., p. 99.

Nemer, M. (1963). Regulation of protein synthesis in the embryogenesis of the sea urchin. Metabolic Control Mechanisms. Natl. Canc. Inst. Monogr. 13: 141-154.

Nijhout, H. F. 1990. Metaphors and the role of genes in development. BioEssays 12: 441-446.

Opitz, J. (1993). Blastogenesis and the primary field in human development. Birth Def. Orig. Art. Series 29 (1): 1-34.

Oyama, S. (1985). The Ontogeny of Information. Cambridge University Press, Cambridge.

Roll-Hansen, N. (1978). Drosophila genetics: A reductionist research program. J. Hist. Biol. 11: 159-210.

Sager, R. and Ryan, F. J. (1961). Cell Heredity. John Wiley and Sons, NY Pp. 321-338.

Sander, K. (1986). The role of genes in ontogenesis: Evolving concepts from 1883-1983 as perceived by an insect embryologist. In A History of Embryology (T. J. Horder, J. A. Witkowski, and C. C. Wylie, editors), Cambridge University Press, Cambridge. Pp. 363-395.

Sapp, J. (1987). Beyond the Gene: Cytoplasmic Inheritance and the Struggle for Authority in Genetics. Oxford University Press. NY.

Saxén, L. (1973). Tissue interactions and teratogenesis. In Pathology of Development. (E. V. Perrin and M. J. Finegold, editors). William and Wilkins, Baltimore, MD. Pp. 31- 51.

Schaffner, K. (1974). Logic of discovery and justification in regulatory genetics. Stud. Hist. Phil. Sci. 4: 349-385.

Scott, R. B. and Bell, E. (1963). Protein synthesis during development: control through messenger RNAs. Science 145: 711-713.

Spemann, H. and Mangold, H. 1924. Induction of embryonic primordia by implantation of organizers from a different species. In B. H. Willier and J. M. Oppenheimer (eds.), Foundations of Experimental Embryology. Hafner, New York, pp. 144-184.

Spiegelman, S. (1948). Differentiation as the controlled production of unique enzymatic patterns. In Danielli and Brown, op. cit, p. 286-325.

Spiegelman, S. (1958) Discussion following the talk of B. Magasanik, In Chemical Basis of Development, Op. cit. p. 491.

Spiegelman, S. and Kamen, M. D. (1946). Genes and nucleoprotein in the synthesis of enzymes. Science 104: 581-584.

Stearns, R. N. and Kostellow, A. B. (1958). Enzyme induction in dissociated embryonic cells. In Chemical Basis of Development, op. cit., pp. 448-457.

Stern, C. (1955). Gene Action, in Analysis of Development (B. H. Willier, P. A. Weiss, and V. Hamburger, eds.), Saunders, Philadelphia, PA. pp. 151- 169.

Sturtevant, A. 1932. The developmental effect of genes. Int. Congress of Genetics. Quoted in Keller, E. F. (in press). Rethinking the meaning of genetic determinism. In Tanner Lectures 1993. University of Utah Press, Salt Lake City.

Twitty, V. C. 1966. Of Scientists and Salamanders. Freeman, San Francisco. p. 39.

Vogel, H. J. (1958). Comment on possible roles of repressors and inducers of enzyme formation in development. In Chemical Basis of Development. op. cit., pp. 479-484.

Waddington, C. H. (1939). An Introduction to Modern Genetics. Allen and Unwin, London. p. 135.

Waddington, C. H. (1939b). Preliminary notes on the development of the wings in normal and mutant strains of Drosophila. Proc. Natl. Acad. Sci. USA 25: 299-307.

Waddington, C. H. (1940). Organisers and Genes. Cambridge University Press, Cambridge.

Waddington, C. H. (1948). The genetic control of development. In Danielli and Brown, Growth in Relation to Differentiation and Morphogenesis. op. cit. p. 145-154.

Waddington, C. H. (1962). New Patterns in Genetics and Development. Columbia University Press. Pp. 14- 36.

Waddington, C. H. (1975). The practical consequences of metaphysical beliefs on a biologist's work: An autobiographical note. In The Evolution of an Evolutionist. Cornell University Press, Ithica. p. 3.

Weiss, P. (1939). Principles of Development. Holt, NY.

Weiss, P. (1962). From cell to molecule. In Molecular Control of Cellular Activity. (J. H. Allen, editor). McGraw-Hill, New York.

Weiss, P. (1968). Dynamics of Development: Experiments and Inferences. Academic Press, NY.

Wilson, E. B. (1925). The Cell in Development and Inheritance. Third edition. Macmillan, NY, p. 1112.

Wimsatt, W. 1984. Reductionist explanation: A functional account. In Sober, E. (ed.) Conceptual Issues in Evolutionary Biology, MIT Press, Cambridge, p. 477-508.