Genetic Analysis 271or many, of the Mendelian differences with which we deal in genetics
as representing changes, chemical in nature, within these individual
genes; or are these differences caused only by decreases and increases
in the number and possibly in size ..., and by changes in mutual
arrangement of genes that themselves remain essentially unchanged,
or even unchangeable? [Muller, 1956, 126]
Once Johannsen established the distinction of phenotype and genotype, it was
manifest that genetic analysis would strive to deduce the nature of the genotype
from observing the phenotype. The chromosomal theory of inheritance provided
a material basis for Mendel’s theory, and indeed, the theory of linkage suggested
the mapping of these bodies. The early virtual maps became very real after the
discovery of the polyteneic chromosomes of Drosophila, and genes, defined by their
(phenotypic) effects, were reference points on these maps. Most geneticist treated
genes as instrumental entities, without committing themselves to the nature of
these entities: They were factors on the chromosomes amenable to genetic analysis.
But in 1922 Hermann J. Muller in a programmatic paper [Muller, 1922], laid
out his notion of the genes as the atoms of heredity, as well as the strategies
that should be adopted to elucidate their structure and function. For Muller
genes were discrete and distinct entities of matter. He explicated their three
essential properties: a) self-replication, orautocatalysis, b) mediating functions and
constructions of living matter, orheterocatalysis, and c) capacity of undergoing rare
changes without loosing their property of self replication, ormutability. Whereas
the first two properties were those immanent to hereditary factors per se, the third
one was essential for organisms that evolve by a Darwinian process of variation
and selection. Although he looked forward for the day when the gene would
be “grounded in a mortar”, that is, analyzed by physico-chemical methods, at
the present time the only way to learn of their structure and function was by
the methods of genetic analysis. Consequently he set out on a research program
focused on the third property of the genes, which was unique to entities of heredity
in a Darwinian system, namely the study of mutations. Following Troland’s [1917]
suggestion, Muller believed for some time that genes were enzymes, or at least
acted like enzymes.
A diametric opposite conception of the genes was that of Richard Goldschmidt,
who maintained a holistic, organismic, top-down conception. According to him
there were no materially discrete and separable physical genes along the chromo-
somes. The chromosomes were integral units of function in a hierarchy of fields of
genetic action, along which one may identify centers of specific functions that if
changed (deleted or rearranged) affected a specific function of the whole. The local
injuries of the chromosomes appeared as if they were discrete genes (Goldschmidt
[1938; 1954; 1955], see also Dietrich [2000]). Although Muller claimed early on
that “no implication as to the physical arrangement of the genes is intended when
the terms ‘linear series’, ‘distance’, etc., are used” [Muller, 1920, 101], he him-
self repeatedly used physical and chemical considerations, arguing that “a little
consideration, however, showed”, that “direct chemical influences of the type in