262 Morphologie, Befruchtung, Teratologie, Cytologie. 
which we are in possession cannot be explained without some such 
material units. The existence of these discrete units, which the 
breeder has been led to assume on purely theoretical grounds, is 
entirely borne out by the observed presence of certain structures in 
the nucleus. Cross breeding experiments have shown that if two 
pure parents differing from one another in one character are mated, 
the offspring behave in a definite and uniform manner; the mem- 
bers of the opposed pair of characters act as independent units. It 
is difficult to escape from the conclusion that these unit characters 
(allelomorphs of Bateson) must be due to material primordia having 
a separate and persistent individuality of their own. Experiments in 
hvbridisation lead to the conclusion that no sexual cell can contain 
more than a single allelomorph from any given pair; this assumption 
is expressed b 3 ^ the term “purity 0 f gametes.” There is only a me- 
chanical mixture of the structual units contained in each of the 
sexual nuclei taking part in the act of fertilisation; the units retain 
their own identity and are again sorted out in different combinations 
when the sexual cells for the next generation are differentiated. The 
act of fertilisation with its concomitant doubling of the chromosomes 
is associated with a correlative process of reduction to one half, — 
meiosis. In the meiotic phase the chromosomes sort themselves into 
pairs, and there are indications that one of each pair is derived 
from the male and one from the female parent, and that it is not 
mere chance which determines which two particular chromosomes 
shall unite to form a pair. The chromosomes themselves cannot be 
the structural units responsible for the characters of the organism, 
on account of their relativety small number. [In the two pure races 
of Pisum sativum and P. arvense there are 18 pairs of characters 
respectively in which the hybrids behave as allelomorphs; but there 
are only 7 chromosomes.] However the chromosomes are themselves 
made up of smaller units, the chromomeres, which are so num- 
merous that they have never been counted in a single chromosome. 
It ma}^ be that these are the units which we are seeking. But if we 
attach to the chromomeres, or to any other still smaller particles, 
the properties of separate character-producing substances, the widely 
entertained view as to the real structural persistance of the chromo¬ 
somes themselves will require some modification. For in order to 
give that complete independence observed to exist between most of 
the allelomorphs, it is clear that any given chromosome must be 
correspondingly indifferent as to which chromomeres enter into its 
composition. In the heterotype pseudochromosomes, each member 
must, however, be composed of homologous primordia contributed 
by the male and female parent respectively. It may well be that 
this is the significance of synapsis, which forms so characteristic a 
feature of the heterotype mitosis. The chromosomes may perhaps 
be compared with the hands that are successively dealt out from a 
pack of cards, each new hand resembling but not being identical 
with those of the preceding deals. What evidence we possess points 
to the conclusion that the two sets of chromosomes, and consequentty 
their chromomeres, remain distinct in all the cell generations up to 
meiosis. Even if it should be found that this distinctness is definitely 
lost in the premeiotic nuclei, it would not weaken the strong evi¬ 
dence in favour of the shuffling of the primordia, and their rearran- 
gement in groups of homologous pairs, at meiosis. The chromomeres 
are to be regarded not as the characters themselves in parvo, but 
as the agents that determine the particular sequence of Chemical 
