ALBERT SZENT-GYORGYI 
the molecule, we will probably find that this is the only molecule 
which can fulfil the specific function of riboflavin. This mole- 
cule has to interact with other molecules built with the same 
precision. Our bodies are built of thousands of such different 
molecules, and chains of molecules. What frightens me is the 
enormous complexity and precision of the system, which has 
now been thrown into relief for the first time by quantum 
mechanics. I find it difficult to believe that such an enormously 
complex system could have been built by blind, random 
mutation. My feeling is that living matter carries, in itself, a 
hitherto undefined principle, a tendency for perfecting itself. 
Whether this principle can be expressed in terms of quantum 
mechanics, I do not know. It is possible that we will have to 
wait for new physical principles to be discovered. Since living 
systems have arisen from inanimate ones, this self-perfecting 
principle may have been present already in the hydrogen atom, 
as the wonderful figures which the frost makes on my window 
in the winter are present, in a way, in every water molecule. 
It may be that life owes its very origin to this principle. 
In reality the situation, in higher organisms, is still much more 
complex. What the body has to do is not simply to improve 
itself, but also to communicate the blueprint of this improve- 
ment to the genetic material, the deoxyribonucleic acid (DNA) 
of sperm and egg cells. The living organism is built from this 
blueprint which has to go through quite different forms of life, 
as shown by the butterfly which has to be, first, a fertilized egg, 
then a caterpillar, then a chrysalis. 
And there is another astounding quality in living systems 
with which I am unable to cope in my thoughts: adaptability. 
Not only are living systems, in a way, perfect, but they are also 
adapted to their surroundings, and if the surroundings change, 
they change too. I am unable to explain this with random 
mutation and I believe that there must be some feedback from 
the periphery to the genetic material. With our present know- 
ledge of coding and communication of the code, it is almost 
possible to give an acceptable theory about the mechanism of 
such a feedback. 
pee 
