398 J. D. BERNAL 



lization, where the studies of Penitz [21] and his school have shown that linking 

 through water molecules plays a predominant role. The amount of water can 

 be very variable, but it would seem as if protein molecules are separated nor- 

 mally only by one or two layers of water molecules and that in the more hydrated 

 cases the large lacunae between the molecules are filled with ordered water 

 molecules as in ice or cr>'ohydrates. Crj'stallization is not, however, a common 

 biological phenomenon and aggregates of this sort are probably of very minor 

 importance in biopoesis. 



The forms of aggregation of nucleic acid do not at the moment seem to be as 

 complex as those of proteins, but this may be simply on account of our smaller 

 knowledge of them. The necessity of packing the purine pyrimidine groups in a 

 quasi-parallel arrangement leads to a very open helix [22]. Indeed it is too wide 

 for a double helix [23], but this seems to be stabilized in the case of DNA by 

 the complementary purine-pyrimidine double hydrogen-bond Unkage, as also 

 is the poly-ribose-adenine, poly-ribose-uridine complex synthesized by Rich 

 [24]. The width and rigidity of the DNA helix would indicate a molecule some 

 1000 Â or more long and forming a parallel hexagonal arrangement [25] and this 

 has also been shown m vivo in sperm heads [26-28]. However, this cannot 

 always be the case, for in the DNA-containing bacteriophages the width of the 

 head is not more than 200 Â, so that some kind of folding or alternative packing 

 must be possible. 



The most complex as well as the most interesting structures based on polymer 

 association are the heterogeneous complexes of which the most studied have 

 been the nucleoproteins, though to understand biological structures much will 

 also have to be learned of the lipoproteins and mucopro teins. At the moment it 

 would appear that there are two different kinds of nucleoprotein : one, occur- 

 ring in sperm, where DNA is linked with small basic protamine molecules ; and 

 the other in viruses containing ribosenucleic acid molecules embedded in a 

 protein shell consisting of large-molecule proteins [25, 28]. In the first case the 

 nucleic acid molecule seems to determine the structure, in the latter case the 

 protein for an almost identical shell can be assembled without nucleic acid. In 

 both cases, however, there can be no doubt that the synthesis of the protein is 

 connected primarily with the nucleic acid, for, in the virus at least, the kind of 

 protein produced in the cell of the infected plant is entirely determined by the 

 kind of nucleic acid introduced. 



It is between particles of the order of magnitude of viruses and microsomes 

 (100-300 Â) that the long-range forces discussed above can most effectively 

 operate. However, it is difficult to follow out their operation in vivo on account 

 of the nvmiber of other particles of approximately the same dimensions usually 

 foimd in cells, and because of the presence of extensive lipid membranes. 

 Theoretically, however, we would expect such forces to be responsible for the 

 slow movements of internal parts of cells in, for instance, chromosome pairing 

 and mitosis, though here there may also be the effect of the formation of fibres 

 by the (g-f) transformation and their subsequent contraction in a way that I 

 have speculated on elsewhere [29]. 



This sketch of the structure and mutual relations of the particles that are 



