Mendel's decision to select such unit characters that would provide insight into the laws of hereditary transmission, whatever their physiological or developmental properties may be, was crucial in the establishment of genetics as an independent discipline of inheritance (Falk, 1995). But when de Vries in 1900 recognized Mendel's work it was because he believed it to contribute to his hypothesis of Intracelluläre Pangenesis, a bottom-up determinist theory of organisms' development and function, as part of the strategic goal of his Die Entstehung der Arten durch Mutation. Indeed, the “unit character” was as much a notion of development as one of inheritance. As conceived by Bateson, unit characters were defined by properties that segregate according to Mendelian rules, and he tried to explicate their role in the physiology and development of the organisms accordingly (see Chapter 3).

Thus, from its inauguration the discipline of genetic research was burdened or blessed with the dialectic tension between the methodological reductionism of hybridists, deducing genes from traits, and the conceptual determinism of morphogenists, inducing traits from genes.

Although geneticists never lost interest in problems of physiology and development (Gilbert, 1998), they did adopt, as a rule, a bottom-up perspective of individual traits rather than a top-down perspective of organisms in their context. However, they needed some empirical handle to examine speculations such as the Weismann–de Vries hypothesis of specific nuclear particles being differentially exported to the cytoplasm of somatic cells.

Johanssen wished to discern the phenotype from the genotype (Johannsen, 1909). “Genes,” however, became what geneticists concluded Mendelian Faktoren had been. Were genes hypothetical constructs, autonomous structural entities or loci of differential functional emphases along an integral chromosome? Or, did all that not matter and genes were simply intervening variables, helpful entities for experimental work (see Chapter 4)?

Strictly speaking, there are no hypothesis-free concepts, and the distinctions between hypothetical constructs and intervening variables that proved very helpful historiographically, are more problematic when one attempts to apply them to personal conceptions. Whether we wish it or not, our concepts “are constructions in thought representing historically an immense amount of intellectual work … the scientific hypothesis does not come after the numerical data but before them” (Woodger, 1967, 366). Morgan, like Johannsen, insisted on presenting genes as intervening variables, with no hypothesis about the nature of the “something.” As late as in his Nobel talk he asserted:

In his classical paper of 1913 on the linear arrangement of sex-linked factors in Drosophila, Sturtevant described the phenotypes of the wings in combinations of genotypes: “The normal wing is RM. The rM wing is known as miniature, the Rm as rudimentary, and the rm as rudimentary-miniature” (Sturtevant, 1913b). The miniature phenotype is produced by a mutation in the rudimentary gene and vice versa, the rudimentary phenotype is produced by a mutation in the miniature gene.

This terminology, unlike the terminology we use today, had been introduced by Morgan: it reflected his sensitivity to the epistemological significance of the notation describing experimental results.

Similarity and variation among living creatures has long been a mystery that has stimulated the classification and organization of life into hierarchical systems, as well as the induction of the differential ranking of individuals in society. Variation and similarity also provided the material for one of the most basic ancient unfolding developments of human culture, animal and plant breeding. The first chapter of Genesis beautifully reflects the ancients' conception of the hierarchical, as well as anthropocentric, catalog of the universe:

Animals and plants were believed to have been generated by God as distinct kinds at the time of Creation, and it was reproduction that maintained the link of similarity in living creatures.

Heredity is, of course, at the heart of any theory of evolution, and more specifically at the heart of the Darwinian hypothesis of evolution by natural selection: it is inherited variability that provides the raw material upon which natural selection can operate. In 1868 Darwin presented a “provisional hypothesis of pangenesis,” in The Variation of Animals and Plants under Domestication. Pangenes, representing all organs, were assembled in the gametes and merged in the zygote to be farmed out to the respective newly formed organs (Darwin, 1868). Although Darwin conceived of “[t]wo distinct elements [that] are included under the term ‘inheritance,’ namely the transmission and the development of characters” (Darwin, 1981 [1871], 279), his hypothesis was essentially a top-down hypothesis of reproduction. As a hypothesis of inheritance it was unacceptable on both theoretical and empirical grounds. Already in 1867, Fleeming Jenkins had shown that in Darwin's blending model of inheritance all variation would be “swamped out” long before it could be established by natural selection (see Hull, 1973, 302–350). Others, like Thomas H. Huxley and, later, Francis Galton, were just as worried about the empirical evidence of the nonheredity of (most of) the intraspecific variation (Falk, 1995, 226).

PANGENES AND FAKTOREN

Contrary to common lore, Mendel's paper was not unknown in the thirty-four years between its publication and its acceptance as significant (see letters to Roberts, in Stern and Sherwood, 1966; Jahn, 1957/58; and Weinstein, 1977).

de Vries and Bateson introduced the hybridist analytic methodology to challenge morphogenists' thinking about Darwinian evolution. Asking to what extent could ecologically related continuous variations of local populations become characteristics of distinct taxonomic strains and eventually species, they adopted not only a reductionist, particulate conception of organisms but also a rather determinist notion of the inherent nature of these unit characters. They established Mendelian genetics on determinist foundations in which Faktoren served as preformed unit characters.

Deterministic notions such as Faktoren becoming unit characters were anathema to both natural historians in the field and embryologists in the laboratory. Richard Woltereck's objective was to provide a rejoinder to what he called the Mendelian teaching following Weismann's and de Vries's conception of the origin of species (Woltereck, 1909). In his “investigations on the change of species, with emphasis on quantitative species-specific difference” he showed that when the conditions under which strains of Daphnia species from different lakes in Germany existed were modified, their morphology changed to simulate that of another acknowledged species. Furthermore, each strain had its specific inherent Norm of Reaction, that is, its characteristic morphological response pattern to variation in environmental conditions. Such specific norms of reaction made it meaningless to predict the properties (the form characteristics) of one strain under one set of conditions from those under another set of conditions or from those of another strain under the same conditions (Woltereck, 1909; see also Falk, 2000a).

Mendelian nomenclature of symbolic representation of genes and their alleles makes it readily amenable to algebraic presentation of population events. Mendel himself was the first to take notice of this when he explained Gärtner's and Kölreuter's observations “that hybrids have a tendency to revert to parental forms” in algebraic terms:

In the last line of the tabulation Mendel generalized the results for the nth generation of selfing that follows mono-hybridization: the ratio of “A : Aa : a” would be 2n−1 : 2 : 2n−1.

Bateson adopted Mendel's hypothesis upon reading de Vries's paper because he believed that it was consistent with his notion of evolution by discontinuous steps. This placed him in opposition to the claim of the mathematical statistician Karl Pearson in favor of Galton's “Law of Ancestral Heredity” of continuous Darwinian evolution (Provine, 1971; and see Chapter 3). In a discussion following a talk by Bateson's colleague Punnett, it was suggested that “a dominant allele, once introduced into a population, would increase in frequency until reaching stability at 0.5, given the usual phenotypic ratio 3 dominant: 1 recessive thereafter” (Provine, 1971, 133–134).

In December 1992 Bertil Daneholt presented the Nobel laureates in physiology and medicine for that year to the assembly of the Karolinska Institute in Stockholm:

The phenomenon of dominance and its meaning accompanied the science of genetics from its initiation. Bateson interpreted it in terms of the Presence and Absence Hypothesis. R. A. Fisher, taking notice of the harmful effects of inbreeding, pointed out that “if we assume that adaptation is the result of selection, the majority of large mutations must be harmful” (Fisher, 1922, 323). Thus a major task of the evolution of dominance would be to equalize the phenotypes of the heterozygotes to that of the homozygote for the wild-type allele (Fisher, 1928). However, doubts were raised about the efficiency of natural selection in carrying out the task (Wright, 1929a). According to Wright's theory of the functional organization of the cell that related the genotype to the phenotype, dominance would be the outcome of interaction:

Haldane (1930) and Muller (1932), following Shull (1909, 415) and Wright, suggested that the amount of product produced by the alleles of a gene varied asymptotically, and that the wild-type alleles in diploids have been selected to function in the relatively saturated segment of the asymptotic gene/product curve (see Chapter 14).

A century ago, “genes” endowed with the instrumental reductionist status of the Mendelian Faktoren were introduced to overcome the notion of the unit character. The gene was an empiric, functionally defined entity. For Johannsen, gene was a word derived from genotype that should express “only the simple imaging” that “through ‘something’ in the gametes a property of the developing organism is conditioned or is, or could be co-determined” (Johannsen, 1909, 124).

Yet, the meaning of the gene, whether as an intervening variable, a heuristic device for the study of inheritance, or a hypothetical construct, an entity in the physico-chemical world, and what kind of such an entity, became a major issue of genetic analysis. From Richard Goldschmidt's “physiological” perspective genes were merely centers of gravity along the integral units of the chromosome. For Hermann Muller genes were discrete structural entities along the chromosomes. Others, like East, Morgan, and Stadler adopted an operational approach of “instrumental reductionism” (Falk, 1986, 141).

Toward the 1940s genetics became a self-confident autonomous discipline. The Nobel Prize awarded to Morgan in 1933 was a symbol of this autonomy and it certainly also added to its self-confidence.

During the first decades of the twentieth century, the methodology of hybridization analysis of discrete traits was established, and a conceptual framework for the mechanics of inheritance was formulated: genes were inherited entities that (within given environmental circumstances) determined the properties of morphological, physiological, as well as behavioral traits of living organisms; the chromosomal theory of inheritance situated the genes along the chromosomes and mapped them; the analysis of mutations combined with cytogenetic observations indicated that there was a material basis for the Mendelian Faktoren and demonstrated the mechanics of their inheritance. Geneticists in Morgan's group and elsewhere now invaded other disciplines and explored them more aggressively: Dobzhansky's Genetics and the Origin of Species (Dobzhansky, 1937) declared his intent to expand into the sphere of evolutionary research.

Already in 1900 Correns suggested that Mendel's “numerical ratio 1:1 strongly suggests that the separation occurs during a nuclear division, the reduction division of Weismann” (Correns, in Stern and Sherwood, 1966, 127). It was, however, only after the establishment of Mendel's laws that cytologists came to share a set of common assumptions which led them to agree on what they saw under the microscope and eventually to accept the link between chromosomes and Mendelian factors. Chromosome continuity and integrity and their specificity were derived from Boveri's experiments; chromosomes' simulation of multi-factorial Mendelian segregation further supported their role in inheritance, as Wilson and his associates indicated in studies of sex-determination in various insect species. But the behavior of chromosomes could not be simply observed; it had to be interpreted: were these just certain other structure-properties that obeyed the Mendelian laws of segregation or were these elements that causally determined the observed Mendelian laws of segregation? The establishment of the distinction between genotype and phenotype, overcoming the preformationist notion of unit character, was crucial for such an examination.

Morgan's discovery of a correlation between the pattern of inheritance of a Mendelian factor of Drosophila and that of its sex-chromosome suggested a causal relationship, but only the finding that the inheritance of several independent Mendelian factors was similarly correlated with sex-chromosomes allowed the hypothesis that the chromosome was the causal mediator of inheritance.

In 1966 George Williams made a heroic attempt to maintain the strict, reductionist approach of the New Synthesis of Darwinian evolution in his Adaptation and Natural Selection: A Critique of Some Current Evolutionary Thought. For Williams,

This uncompromising bottom-up doctrine opposed and rejected “certain of the recently advocated qualifications and additions to the theory of natural selection, such as genetic assimilation, group selection, and cumulative progress in adaptive evolution” (Williams, 1974 [1966], 4).

When methods for in vitro hybridization of polynucleotide molecules were developed by Doty and Marmur and their colleagues (Marmur and Lane, 1960), it opened new vistas for genetic research. These new methods soon transformed and extended much of genetic research to genetic analysis of hybridization at the polynucleotide level. Such notions had been anticipated by Crick regarding the genetics and taxonomy of proteins (1958, 142): “Biologists should realize that before long we shall have a subject which might be called ‘protein taxonomy’ – the study of the amino acid sequences of the proteins of an organism and the comparison of them between species.” The beginnings of these analyses, developmental (pathological) on the one hand and taxonomic on the other, were already evident from the studies of changes of single amino acids in hemoglobin molecules (Ingram, 1957, 1963).

Hoyer, Bolton, and McCarthy isolated RNA complementary to DNA, based on the demonstration that specific cellular RNA could be hybridized with heat-denatured DNA from the same cells. The hybrids were isolated by cesium chloride density gradient centrifugation (Bolton and McCarthy, 1962). These researchers examined primarily bacterial species relationships, “where there exists only the faintest paleontological record and the simplest of all ontogenetic processes.”