922 
acter being dependent on several factors 
and each factor affecting a number of 
parts and even qualities of the plant. 
Even the most thoroughgoing Mendelian 
must admit that a unit character made up 
of four fractions or assumed to be depend- 
ent for its realization on four factors has 
lost something of its unity. What the stu- 
dent of the germ plasm wants to know, of 
course, is the nature of these fractions or 
the ultimate elements in heredity, what- 
ever they may be. The breeder may per- 
haps properly, as Baur does, relegate all 
questions as to the nature of the represen- 
tation of the hereditary qualities of the 
many-celled organism in the egg to the 
future as outside the scope of his immedi- 
ate experiments. The cytologist dealing 
directly with the germ plasm in the chro- 
mosomes is confronted directly with the 
question of their ultimate constitution and 
must attempt to connect any discoverable 
units in the make-up of the adult in some 
way with the structure and properties of 
the germ plasm itself. It is obvious that 
visible characters which may arise in more 
than one way or that require the combina- 
tion of from two to several factors for their 
production can hardly be represented in 
the germ plasm by the pangens, determi- 
nants or other corpuscular units of the 
older theories. To call a visible character 
which depends on four hereditary factors 
for its production a unit character is cer- 
fainly not conducive to clearness of 
thought. To attempt to find fixed units 
in the maze of fluctuating colors, forms 
and physiological processes of multicellu- 
lar plants looked upon as wholes is per- 
haps a hopeless task. Much more is it in- 
conceivable that such diffuse fluctuating 
characteristics are represented by specific 
corpuscles in the germ plasm. 
We have noted that Mendelian breeding, 
emphasizing as it does the existence of 
SCIENCE 
[N.S. Vou. XXXV. No. 911 
characteristics belonging to the organism 
as a whole and their couplings and repul- 
sions in inheritance may have a most im- 
portant result in the elimination of the 
last trace of the doctrine of preformation 
from our conceptions of the germ plasm. 
It seems to me also probable that the con- 
sistent study of the so-called unit char- 
acters is tending rapidly to the overthrow 
of all corpuscular theories of heredity, and 
that with a proper understanding and in- 
terpretation of the Mendelian factors we 
may finally be freed from these confusing 
molecular chemical analogies in the study 
of the germ plasm. 
There are certain broad inconsistencies 
in the doctrine that the characters of the 
whole organism as such are represented in 
any fashion by units of the germ plasm, 
which should always be borne in mind. 
Most conspicuous of these is the fact that 
there is no proportion between either the 
number of the chromosomes or their mass 
and the complexity of the organism to 
which they belong. The simplest algz and 
fungi may have as many chromosomes and, 
proportionally to the size of the cell and 
nucleus, as large chromosomes as some 
flowering plants. Allowing for a large 
amount of possible ultra-microscopic or- 
ganization, such disproportions are not 
consistent with any corpuscular theory of 
heredity. It may well be that just as 
many of the so-called unit characters of 
Mendel and De Vries relate to the diffuse 
properties of the organism, as a whole, so 
the hereditary factors representing them 
depend on diffuse qualities of the cells as 
wholes in their interactions with each other. 
Such a view is not inconsistent with the 
doctrine that the chromosomes are the 
physical basis of heredity. We should per- 
haps, with Hertwig, more clearly distin- 
euish between the heredity which deter- 
mines the characters of the cells, epider- 
