50 



TEMPORAL ORGANIZATION IN CELLS 



and reliability will undoubtedly make an important contribution to the 

 understanding of cell behaviour. 



One extremely interesting recent development which will certainly have 

 great importance for the analysis of biological systems, is the extension by 

 Winograd and Cowan (1963) of the techniques and theorems of information 

 theory to computation by automata. In the context of our present study, the 

 work of these investigators might be applied to the question of obtaining some 

 information about the motion or the distribution of the steady state values 

 (Phli) which are simply assumed to exist in our theory of biochemical control 

 mechanisms. Such a "higher-level" analysis would perhaps involve regarding 

 these steady state quantities as probabilities of transmission of different mRNA 

 and protein species in some kind of developmental "message". The network 

 of cellular interactions must function so as to compute and decode reliably the 

 programme of embryonic development which is partly given and partly evolves 

 in the system. These procedures enjoy a degree of generality not shared by the 

 present theory in that they can be applied to systems wherein the interactions 

 between components need not be well defined, and can be of either an inhi- 

 bitory or excitatory (repressive or inductive) nature. However, the price to 

 be paid for generality is a loss in predictive power, and the specific results 

 which can be obtained by the present state of automata theory are fairly 

 restricted. Furthermore, they tend to be algebraic or logical rather than 

 dynamic, and it is the latter in which we are most interested in this work. It 

 is also true that automata theory at the moment is most appropriate for 

 situations where the pattern of interactions is least defined and most complex; 

 for example, theorems on reliability apply to systems with a high level of 

 redundancy. This condition of high redundancy will no doubt be satisfied in 

 many biological systems; but in the case of biochemical control systems in 

 cells there is apparently a lot of precise, deterministic structure in the pattern 

 of molecular interactions. Such a detailed microscopic structure is best 



(26) 



