1892.] Heredity and the Germ- Cells. 663 
of each division figure are paternal and two are maternal. In 
favor of this hypothesis we may place the following facts: 
1st, that there are an even number of chromasome rods in all 
cells; 2d, that the number is constant throughout all the sub- 
sequent changes in the tissues; 3d, that the number is fixed 
for each species or variety; 4th, that the number is the same 
in each sex. 
Against this replacement hypothesis we must consider the 
extreme complexity of the division process, and the long-rest- 
ing, or thread stage, in which the chromatin lies in a confused 
coil. Further, Hertwig argues that if the elements are dis- 
tinct we should find some evidence that the maternal or pater- 
nal part is atrophied or replaced, or excluded from the nucleus, 
for both parts cannot share alike in the control of the cell. 
These are Hertwig’s grounds for supporting the “ verschmel- 
sungstheorie,” or fusion theory, also advocated by Waldeyer, 
to the effect that by the complete union of the maternal and 
paternal substance a new product is formed; in this fusion the 
law of prepotency may come into play, causing one or other 
of the parental tendencies to predominate, or there may be an 
even redistribution, whereby, as expressed by Hensen, “the 
hereditary substance of the son is not that of the father plus 
that of the mother, but is his own, with a new hereditary form 
resulting from the combination.” 
While suspending judgment between these two views as to 
the separation or fusion of the chromatin, we may appeal to 
the external phenomena of heredity for light upon the proba- 
bilities in the question. First, I refer to the very decided 
opinion of Francis Galton in regard to particulate inheritance ; 
he is so impressed with the fact that we are made up bit by bit 
of separate structures derived from different ancestors that he 
has even suggested that the skin of the mulatto may represent 
not a fusion of white and black but an excessively fine mosaic 
in which the colors are so distributed as to give the appear- 
ance of blending. We do sometimes observe patches of color 
as evidence of uneven distribution. As Galton distinguishes 
two types of structures with reference to inheritance, viz., those 
which blend and those which do not blend, we might corre- 
