190 GENETICS IN RELATION TO AGRICULTURE 
definite with respect to the factors involved. Such an assumption is 
obviously an hypothesis of extensive multiple allelomorphism in which the 
members of the system Aj, As, Az, . . . A, when in contact with each 
other, react to form new allelomorphs. On the basis of a chemical con- 
stitution for the locus in the chromatin material, such an hypothesis would 
appear not improbable, for if the nucleus of this locus were identical in all 
the allelomorphs involving it and the changes in it were changes which took 
place around the fringe of the molecule, inside and end chains perhaps, 
then there appears to be no good reason for believing that two such similar 
allelomorphs when in intimate association with each other should not 
interact to form intermediate factors. While the very existence of the 
second and third categories is not generally accepted by geneticists, it 
must be admitted that, provided the assumptions of factorial alterability 
be accepted, they can explain the known facts of size inheritance. 
That a factor as such may vary and that selections of variations in the 
expression of such a factor may permanently alter its expression in any 
desired direction, Castle has endeavored to show in extensive selection - 
studies with hooded rats. The selections were made in both plus and 
minus directions and in each case selection did establish a permanent 
variation in the direction in which it was made. The condensed results 
of these selection experiments involving over 30,000 individuals are given 
in Tables XXXIV and XXXV. These results show very clearly that 
selection has definitely changed the mode around which the pattern 
fluctuates and in opposite directions in the two series. Since the hooded 
pattern has been shown to be a simple Mendelian recessive to self color, 
does this evidence prove that selection has modified the factor for hooded 
pattern? 
Those who hold to the multiple factor idea of size inheritance contend 
that it does not, and for several reasons. In the first place the difference 
between hooded and self-colored rats, while in the main due to a single 
factor difference, may at the same time involve other minor differences in 
a number of factors which influence the eatent of pigmentation when the 
factor for hooded pattern 1s present. According to this view selection 
would result in attaining a homozygous condition for certain of these 
modifiers and, therefore, in increasing or decreasing the area of pigmen- 
tation in the direction of selection. There may be a number of such 
modifiers and others may arise from time to time by mutation. It is 
interesting to note that Castle records the appearance of two such 
mutant individuals. These mutant individuals when tested with each 
other and the forms from which they arose, displayed a type of inher- 
itance which indicated that their origin involved a single factor differ- 
ence for extent of pigmentation from the parental group. Might not 
other mutations have arisen which, on account of their lesser magnitude 
Digitized by Microsoft® 
