levels, from that of the bacterial plasmids to sexual recombination in 

 the classical way; even the typical microbial cell may itself largely be 

 a symbiotic union of once-independent organisms. Some such hier- 

 archical architecture is probably essential, even a rare process which 

 increases the probability that successful changes can come to domi- 

 nate. The slow growth of molecular tapes in solution — whether one 

 imagines an organic ocean or a more plausible multitude of enriched 

 tide pools — is bound to be speeded up by any process which tends 

 to sequester reactants in the right way. 



The logical simplicity of the replicating single molecule is evi- 

 dent. On the other hand, the complex system we now see requires 

 the participation, even in vitro, of an energy source and a number of 

 specific auxiliary enzymes. In the cell of course much more is invari- 

 ably present, from the multimolecular mechanism of the transcrip- 

 tion apparatus to the enclosing membranes which maintain the 

 concentrations and the integrity of every cell. The molecular cou- 

 pling described above is a minimal early step. 



It is plausible to extend the idea of the single molecule. The 

 simplest extension is to a system of a few molecules, which seems 

 closer to the working system of life today. The difficulty is still in 

 the coupling within any simple molecular system. 



The main logical point seems to be to extend the participation 

 of the environment, in order to restrict the replicating event to a 

 single protogene itself. To this end the free energy is thought to be 

 provided by a medium rich in needed building blocks, stored there 

 by nonbiological processes that produced the molecules required. It 

 is not difficult either to imagine the presence of some nonspecific 

 catalyst, say, a metal ion, which might assist the coupling sought 

 between the major polymers. 



The next step beyond the single molecule self-copying in an 

 organically rich and slightly metallized medium is the idea of a non- 

 specific substrate, some solid mineral surfaces where adsorption 

 might locally increase concentration, provide nonspecific catalysis, 

 and allow the use of rare components. At this point of hierarchi- 

 cal steps of unification, perhaps on the model of the simplest one, 

 if A can make not itself but can make B, while B in turn makes 

 A, the single union is self-duplicating. Once the possible role of 

 substrate is admitted, the scheme can go even further. We shall 

 discuss in more specific terms in chapter VI the conjectures that 



