918 
pangen characters, such as those for the 
production of specific alkaloids, oils, tan- 
nin, etc., involve the same difficulties, and 
this is no less true for morphogenetic 
pangens of leaf form, ete. To speak with 
Klebs, such pangens should be potentiali- 
ties rather than material granules. 
The failure of the experimental breed- 
ers to characterize more exactly the hypo- 
thetical units, genes, ete., in the germ plasm 
which represent their factors is not all due 
to a desire for absolute freedom from prej- 
udice in matters of theory, but rather to an 
inability to identify the behavior of their 
factors with that of any such units in the 
germ plasm as the older theories have as- 
sumed. Johannsen gives a name to the 
units of which his germ plasm is composed, 
the genes, but he defines genes and geno- 
type strictly from the visible behavior of 
the different characters of the many-celled 
plant. The genotype is the specific germ 
plasm, each of the genes represents a unit 
character. How they are situated or re- 
lated in the egg is unspecified, though it is 
suggested they probably have only a chem- 
ical constitution. 
The Mendelian unit characters are, as 
noted, most typically generalized qualities 
of the plant or organ as physiological 
wholes. They are very diverse in their 
character and there is little attempt as 
yet to classify them. We have quantita- 
tive characters of weight and measure in 
fruits, stems, etc., superficial factors of 
color, which palpably depend on nothing 
more fundamental than a slight change 
in the degree of oxidation of a by-product 
of the protoplasm or a variation in the 
alkalinity or acidity of the cell sap. 
We have factors for annual and bien- 
nial habit, as well as the more fundamental 
factors for form and tissue differentiation 
which are essential to the every-day exist- 
ence of the organism. These heterogeneous 
SCIENCE 
[N.S. Vou. XXXV. No. 911 
factors show couplings, repulsions, etc., 
without regard to any, at present, explain- 
able relationship between them which 
would make it possible to associate them 
with any definite spatial distribution of 
anlagen in the germ cells. 
It may be found to be one of the most 
valuable results of Mendelian experimen- 
tation that it has helped to destroy the last 
vestiges of preformationist ideas which in- 
hered in the prearrangement of the hered- 
itary corpuscles assumed in the theories of 
Weismann. It is certainly impossible to 
imagine where a corpuscle should be 
placed in the egg or what it should do to 
change a plant from an annual to a bien- 
nial, or a crenate to a serrate leaf. 
It is sufficiently clear that the results of 
Mendelian and mutational breeding from 
the standpoint of the kind of unit factors 
observed and their behavior in cases of so- 
called coupling and repulsion are opposed 
to the doctrine of a spatially organized 
germ plasm made up of corpuscular units. 
The doctrine of the fixity of the unit char- 
acters and their segregation as pure ele- 
ments seems still, however, to harmonize 
well with the conception of pangens, per- 
haps not so definitely related to each other, 
spatially, in the germ plasm. If there are 
fixed elements, unit characters, which are 
transmitted in reproduction, this is cer- 
tainly strong ground for the assumption 
of the existence of corresponding corpuscu- 
lar units in the germ plasm. The experi- 
ence of experimental breeders with the doc- 
trine of segregation must be examined 
critically as to the evidence it gives on this 
point. It is a fundamental assumption of 
Mendelism that the characteristics of the 
parents behave in fertilization, whether 
hybrid or normal, as unit characters, and 
that hybridization and the phenomena of 
segregation which follow give the best pos- 
sible means of recognizing and identifying 
