efore Darwin, most biologists adheredto a platonic model of nature. Thisimplied that the biological realm con-sisted of a finite set of essentially immutablenatural forms that, like inorganic formssuch as atoms or crystals, are an intrinsicpart of the eternal order of the world. Justas, today, we account for the form of atomsand crystals by a set of physical laws or‘constructional rules’, so pre-darwinianbiologists sought to account for the originof biological forms in terms of a set ofgenerative physical laws often referred to asthe ‘laws of form’.For many biologists today, platonicbiology is an anachronism irretrievably laidto rest, and the idea that biological formsmight be intrinsic features of nature generat-ed by physical laws is treated with increduli-ty. However, recent advances in proteinchemistry suggest that at least one set ofbiological forms — the basic protein folds —is determined by physical laws similar tothose giving rise to crystals and atoms. Theygive every appearance of being invariantplatonic forms of precisely the type that thepre-darwinian biologists were seeking. Protein folds, the basic constructionalunits of proteins, each consist of a foldedchain of between 80 and 200 amino acids.Some proteins consist of a single fold, butmost are a combination of two or more.During the 1970s, as the three-dimensionalstructure of an increasing number of foldswas determined, it became apparent thatthe folds could be classified into a finitenumber of distinct structural families con-taining a number of closely related forms.The fact that protein folds could be classifiedin this manner provided the first line ofevidence that the folds might be naturalforms.Further evidence that the folds doindeed represent a finite set of naturalforms is provided by detailed structuralstudies carried out over the past two decadeswhich have revealed that the structure ofthe folds can be accounted for by whatamounts to a set of ‘constructional rules’governing the way that the various sec-ondary structural motifs, such as a-helicesand b-sheets, can be combined and packedinto compact three-dimensional structures.One is inevitably reminded of the atom-building rules governing the assembly ofsubatomic particles into the 92 atoms of theperiodic table. Consideration of these ‘constructionallaws’ suggests that the total number ofpermissible folds is bound to be restrictedto a very small number — about 4,000,according to one estimate. Confirmationthat this is probably so is provided by a differ-ent type of estimate, based on the discoveryrate of new folds. Using this method,Cyrus Chothia of Britain’s Medical ResearchCouncil estimated that the total number offolds utilized by living organisms may not bemore than 1,000. Subsequent estimates havegiven figures of between 500 and 1,000.Whatever the final figure, the fact thatthe total number of folds represents a tinystable fraction of all possible polypeptideconformations, determined by the laws ofphysics, reinforces the notion that the folds,like atoms, represent a finite set of built-innatural forms.The robustness of the folds offers anoth-er clue. The fact that the folds can retaintheir native conformations in the face ofmultiple different sorts of short-term defor-mations caused by the molecular turbulenceof the cell, and in the face of extensive, long-term evolutionary changes in their amino-acid sequences, is precisely what would beexpected if they are natural forms, specifiedby physical law. Again, the fact that thesame fold can be specified by many differ-ent, apparently unrelated amino-acidsequences, suggesting multiple separate dis-coveries during the course of evolution, isfurther evidence that the folds are intrinsicfeatures of the order of nature. Finally, thefact that in many cases the same fold isadapted to very different biochemical func-tions is precisely what would be expected ifprotein functions are secondary adaptationsof a set of primary, immutable, naturalforms. If forms as complex as the protein foldsare intrinsic features of nature, might someof the higher architecture of life also be deter-mined by physical law? The robustness ofcertain cytoplasmic forms, for example thespindle apparatus and the cell form of ciliateprotozoans such as