272 Raphael Falk
question could hardly extended over distances so relatively vast from the chemical
standpoint, as are those here in question” [Muller, 1938, 587].
It is important to remember that the basic empirical test for differentiating genes
as distinct units was afunctionaltest, the “complementation test”: Two (reces-
sive) mutations were confronted with each other in the same cell in heterozygotes
(or heterokaryons in fungi, merozygotes in bacteria or multi-infection in viruses).
If the phenotype of the compound heterozygote was normal (“wild type”), the
mutations affected different genes; if, however, the compound heterozygote was
phenotypically mutant, both mutations affected in the same gene (even if their
own phenotypes were different). Thestructuralproperties derived from such a
defined entity were the ones disputed. One piece of evidence against mutations
being all deletions or rearrangements of more integrative entity, as claimed by
Goldschmidt, was the observation that the same gene could mutate to several alle-
les, each affecting the same characteristic but at different intensity, like mutants of
the genewhitethe alleles of which varied for eye color from white to apricot, buff,
eosin, coffee, coral, ivory, blood, etc. (denotedw, wa,wbf,we,wcf,wco,wi,wbl,re-
spectively). This, however, could be settled if one adopted Lewis Stadler’s [1954]
operationalmodel of the gene as a linear series of points within a continuum,
rather than an indivisible atom of inheritance, so that affecting (deleting) any one
may cause functional disturbances, which may be observed as mutations of the or-
ganism. Muller found further support for the “atomic” gene concept in “position
effect”, where it was shown that not a deletion of the gene but, at least in spe-
cific cases, its rearrangement with respect to its neighbors, was responsible for the
mutated effect [Raffel and Muller, 1940]. According to him, these “apparent muta-
tions of genes located near points of chromosome breakage are only changes in gene
functioning conditioned by an alternations of gene grouping” [Muller, 1938, 587]:
Changing function by changing neighbors, has meaning only if neighbors do exist.
Goldschmidt, on the other hand, claimed position-effects to provide strong evi-
dence against the discreteness of the genes; it supports functional integrity rather
than structural discreteness. Rearrangements that resulted in position-effect par-
tial inactivation of genes (such aswM^4 , white-mottled eye color) yielded upon
X-ray irradiation new alleles of various grades of inactivation or reactivation, de-
pending on the amount of heterochromatin (condensed chromosome segments, of
poor genetic activity; contrary to euchromatic segments) present next to the site
of the original gene [Panshin, 1935; 1936]. Muller considered this as evidence for
the change in the configuration of genes due to the variation in their milieu that
resulted in changed gene activity, analogous to that known for enzyme activity.
However, Muller’s argument for genes that may mutate (and not just be deleted)
was ana priorione, anchored in his Darwinian notion of evolution by small dis-
crete changes, according to which “it is not inheritance and variation which bring
about evolution, but the inheritance of variation, and this in turn is due to the
general principle of gene construction” [Muller, 1922, 35]. Genes, not whole chro-
mosomes must be the units of evolution. He found further evidence for discrete
genes being the units of evolution in the fact that “wild type” alleles were as a rule