Mutations and Evolation. 69 
factors and hence round, could come about through the detachment 
of the determiner for triangular capsule from the end of one 
chromosome and its attachment to the end of the next, seems more 
likelyand more in accord with our present knowledge. The difficulty 
with it as a general explanation is that one cannot suppose that all 
duplicated factors have been conveniently located on the end of a 
chromosome, especially where a number have been described for 
the same organism, as is the case with wheat, maize and tobacco. 
The most likely hypothesis of the origin of most duplicate factors 
is then the independent origin of each through a chemical alteration 
of a locus in a different chromosome. 
However, the recent breeding experiments with Drosophila 
have disclosed cases (Bridges 1917) of duplication in which the 
genetic behaviour is as though a group of genes from the middle of 
one X-chromosome has become attached to the end of the other 
X-chromosome in a female. Another exceptional case, discovered 
by Bridges and reported by Morgan (1919), is explained on the 
assumption that a piece from the second chromosome has become 
attached to the middle of the third chromosome. In the resulting 
race, when these chromosomes separate and recombine in reduction 
and fertilization, zygotes which receive the deficient second 
chromosome fail to develop unless they also receive the third 
chromosome with the additional (duplicate) piece. We are 
therefore at liberty to suppose that a redistribution of certain 
chromatin elements, rather than a fresh transformation of a new 
determiner has taken place in certain instances. But such cases 
will usually involve a group of factors simultaneously rather than 
a single one. 
Recessive Mendeliau Factor Mutations. 
If we turn now to CEnothera gigas nanella as the type of a 
recessive Mendelian factor mutation, the manner of its origin 
seems clear from its hereditary behaviour. De Vries (1915) has 
brought together the evidence concerning its behaviour. CE. gigas 
produces this dwarf in 1—2% of its offspring, i.e., as mutations. 
But certain individuals of gigas are known, from observations of 
Schouten, Gates and de Vries, to produce as many as 18% of dwarfs. 
Such individuals are evidently heterozygous, arising from the union 
of a normal gigas germ cell with one which has mutated so as to 
carry the dwarf factor instead of that for tallness. Theoretically 
they should give 25% of dwarf offspring but the number is reduced 
by their lesser viability. This phenomenon of producing in the 
