302 
tions, even when they happen to 
change in such a way as to initiate 
new reactions. It seems likely that 
the most favorable condition for the 
production of such new functions is 
one in which some of the usual genes 
are present in duplicate. Cases of hy- 
brids with doubled chromosome num- 
bers, such as the cottons discussed 
above, furnish such an opportunity, 
for in these cases there is a whole extra 
set of genes, whereas a single set is all 
that is needed to carry on the func- 
tions normal to such an organism. 
There is evidence, particularly in the 
case of the tobacco plant (2), that such 
hybrids gradually lose one set of 
genes—or rather parts of the set derived 
from one original parent, and other 
parts of the set derived from the other 
parent. Nevertheless, a considerable 
period is present in which duplicate 
genes are present and are available for 
the trying out of new experiments 
without the loss of any of the estab- 
lished and useful reactions. 
This cannot be the only answer to 
the problem, for hybrids of this kind, 
while rather frequent in the higher 
plants, are very rare in animals, and 
animals also develop new functions in 
the course of evolution. It may be 
suggested that here also there is a 
source of new genes. It happens that 
in certain of the Diptera, including 
Drosophila, the chromosomes in the 
salivary gland cells are unusually large, 
and permit a study of the details of 
their structure to a degree of refine- 
ment nowhere else attainable. One 
result of such a study is the discovery 
of “repeats’—small sections present 
in duplicate (1). It is not known how 
widespread the “‘repeats” are, for they 
cannot be detected in most material. 
Little is known as yet about the gene 
content of such “repeats.”” However, 
it seems probable that they occur in 
most chromosomes, and that they do 
in fact represent duplications of genes. 
These “repeats” may, therefore, be a 
source of extra genes that are not 
needed for the maintenance of existing 
ANNUAL REPORT SMITHSONIAN INSTITUTION, 1948 
functions, and that may therefore be 
used by the organism in trying out 
new kinds of reactions. 
One thing that is definitely known 
about genes is that they reproduce 
themselves. At least once per cell 
division, on the average, each gene 
somehow conditions the formation of 
a copy of itself. On the view that 
gene specificity is due to shape rather 
than to gross chemical composition, 
the simplest assumption is that the 
new gene is moulded about the old 
one. This process may be pictured 
most easily if each gene is thought of 
as being only one layer thick, so that 
determination of the shape of one 
face automatically fixes the shape of 
the opposite one also. 
There are in Drosophila some dozens 
of successive cell divisions between the 
egg of one generation and that of the 
next. It isnot known how many gen- 
erations of individuals separate the 
members of one species from those of 
another species, but hundreds of thou- 
sands is clearly a conservative estimate. 
Multiplying these two numbers to- 
gether, we find that like genes in dis- 
tinct species must have resulted from 
some millions of successive copyings 
during the long period since they had 
a common model. Whatever be the 
process of gene reproduction, it is evi- 
dently an extraordinarily precise one. 
REFERENCES 
This list is intended only as a series of 
suggestions for those who may be interested 
in following up particular subjects. 
1. Brinces, C. B. 
1935. Salivary chromosome maps. 
Journ. Hered. vol. 26, pp. 
60-64. 
2. CLausen, R. E. 
1941. Polyploidy in Nicotiana. Amer. 
Nat., vol. 75, pp. 291-306. 
3. EMERSON, S. 
1945. Genetics as a tool for studying 
gene structure. Ann. Mis- 
souri Bot. Garden, vol. 32, 
pp. 243-249. 
4. Harwanp, S. C. 
1936. The genetical conception of the 
species. Biol. Rev., vol. 11, 
pp. 83-112. 
