JUNE 14, 1912] 
ogous chromosomes are paired in synapsis, 
and the prophases of the reduction divi- 
sion. The evidence from cases of chromo- 
somes of unequal length in both animals 
and plants seems convincing on this point. 
The determination of the relative position 
of the parental elements in the vegetative 
cell generations would go far to settle the 
vexed question of whether this pairing is 
side by side or end to end. Without going 
into the evidence on this point, so many 
times reviewed in recent years, I may ex- 
press my opinion that it favors the side-by- 
side conjugation and, further, indicates 
that the union of the two parental germ 
plasms is, in many cases at least, a very 
intimate one, so that in the pachyneme 
spireme the visible identity of the two pa- 
rental elements completely disappears. 
The discovery that this intimate union of 
the germ plasms comes at the close of the 
F, generation in the preparation of the 
germ cells for the F’, generation forms per- 
haps one of the closest points of contact 
between the results of cytological study 
and experimental breeding. The long- 
known relative constancy of the first hy- 
brid generation contrasted with the break- 
ing up in the succeeding generations has 
here its counterpart in the relations of the 
germ plasms in the reproductive cells. 
It seems plain to me also that the he- 
havior of the chromosomes in thus uniting 
so intimately that their visible identity is 
lost in the pachyneme spirem, is strongly 
opposed to the conception of universal and 
absolute gametic purity and unit charac- 
ters. It is a very obvious suggestion that 
the elements of the gametes should not be 
pure after this union. This obvious sug- 
gestion from cytology may clear up the 
behavior of the offspring from many cross- 
ings better than the assumption of more 
unit factors. The behavior of the parental 
chromosomes in synapsis and the follow- 
SCIENCE 
915 
ing stages is well calculated to provide for 
just such fluctuating variability as well as 
a certain degree of stability as the breeder 
unbiased by Mendelian preconceptions 
finds. It is quite possible that in some 
cases pairs of parental characters may sep- 
arate or interchange without a trace of 
mutual influence, but the close union found 
in the synaptic knot and the succeeding 
spirem certainly seems adapted to provide 
the opportunity for a vast amount of mod- 
ification and interaction between the pa- 
rental germ plasms. 
At this stage again we are confronted 
in very concrete form with the alterna- 
tives of a chemical and mechanical organ- 
ization of the chromosomes and the diffi- 
culties involved in the corpuscular theories 
become very conspicuous. There is no 
visible provision in synapsis for maintain- 
ing any such space relations between the de- 
terminants as Weismann’s theory requires, 
and if we consider the case of the more 
vaguely arranged pangens of De Vries, it 
is still at least very possible that chemical 
reactions might occur between these mi- 
nute proteid masses so intimately associ- 
ated as they are. 
In the phenomena of reduction, on the 
one hand, and segregation, on the other, 
the work of the cytologist and experimental 
breeder finds a most intimate point of con- 
tact, and tne results of studies on these 
phenomena from both standpoints must 
have the most profound and far-reaching 
effects on our theories of heredity. The 
facts of synapsis are even more opposed 
to complexity of organization in the germ 
plasm than are the facts of nuclear division 
which have been so much emphasized. 
To understand the present point of view 
of cytologists and breeders more clearly 
we must briefly examine the current cor- 
puscular conceptions of the germ plasm. 
Detto’s analysis and criticism of these 
