The Problem of Stages in Biopoeds 41 



then would it seem to mark a resting stage of a process which in full activity 

 does not, and perhaps could not, occur in such an enclosed state. This we cannot 

 know until their mechanism of reproduction has been observed. 



However that may be, these small bodies 100-300 A in dimensions appear as 

 the simplest macromolecular sub-units or organelles which form the popula- 

 tion of cells or existing organisms. It would seem not unreasonable to imagine 

 that they antedate cells themselves and may have first appeared as inhabitants 

 of either the diffuse pools of metaboUzing material or the coacervate drops 

 already described. The protein-covered resting stage that we observe to-day 

 in viruses may indeed in the first place have served to protect them when the 

 rest of the system dried up. However, neither the nucleoprotein organelles nor 

 the viruses can be considered as organisms in the full sense, or even by themselves 

 as proto-organisms, because of their extremely limited and speciahzed meta- 

 bolisms. 



More elaborate organelles such as plastids, mitochondria and Golgi bodies, 

 as well as some DNA viruses, possess another element which seems to mark 

 them off as later in biopoesis. This is the presence of membranes, seeming to 

 consist of phosphoHpid, usually appearing in double or multiple layers con- 

 taining hydrocarbon chains of 16-20 atoms long. The present biosynthesis of 

 lipids is much too compHcated for it to have been their original synthesis. 



Nevertheless they probably originated, then as now, from a condensation of 

 two-carbon acids which in turn may be derived from sugars. I am still inclined 

 to put their first appearance as somewhat late, but convincing evidence for this 

 is lacking and this gap in our knowledge seriously hinders the building of a 

 consistent scheme of biopoesis. 



Because a hydrocarbon chain of more than a few atoms is hydrophobic, with 

 the first synthesis of hpids there appears a foreign element in an aggregate of 

 water-soluble molecules. Such lipid molecules are driven together to form 

 micelles which if they are long and uniform take the shape of bimolecular ribbons 

 or sheets. Though such micelles are not polymers in the chemical sense of being 

 held together by covalent links, they play the same role in two dimensions as 

 linear polymers do in one, though they are far more easily disaggregated and 

 reaggregated. The function of lipid sheets in cells seems to be twofold: first as 

 divisions between parts of a cell ; and second as the basis of attachment of other 

 molecules, either active enzymes as lipoproteins or passive Hpid-based pro- 

 tein membranes. This first function is an intrinsic consequence of the packing 

 of long-chain hydrophobic molecules. Such sheets will automatically set them- 

 selves across concentration gradients and make it possible by closing up to main- 

 tain vesicles with a different chemical composition from their surroundings. 

 The membrane-limited drop differs from the coacervate drop in holding up the 

 diffusion of molecules and thus allows for much more intra-organismal spe- 

 cialization. 



Double Hpid membranes, such as those of soap bubble but with the hydrophilic 

 side out, are intrinsically stable. They have been demonstrated in the electron 

 microscope round such elementary structures as animal (DNA) viruses and 

 bacteriophages and, although its structure seems more complicated, the cell 



