966 THE ANIMAL VIRUSES 



Valley fever. A preliminary hyperplasia may be succeeded, as in the poxes, by 

 necrosis. 



The Nature of Viruses. — Up till a few years ago the evidence was becoming 

 increasingly stronger that the animal viruses were essentially minute micro- 

 organisms. Microscopical, filtration and centrifugal observations left little doubt 

 that they consisted of relatively large particles which, in each virus, were of fairly 

 uniform size. The agreement between the results of these three methods 

 of examination was, in fact, so good as to make it necessary to conclude that, 

 if the visible coccoid bodies were not the infective units of the virus, then the 

 virus must consist of particles of the same size which, for some reason or other, 

 were invisible either by direct observation or by ultra-violet photography (Dale 

 1935). In their morphological appearance, in their formation of discrete colonies 

 under suitable conditions — as on the chorio-allantoic membrane of the developing 

 chick embryo — in their coinplex antigenic structure, in their ability to stimulate 

 the production of different types of antibody, in their pathogenicity and their 

 selective tissue localization, in their capacity for variation, in their neutralizability 

 by specific antiserum, and in their power to give rise to immunity to fresh infection, 

 some of the better-studied viruses were, apart from their small size and their 

 failure to grow on lifeless media, indistinguishable from ordinary bacteria. 



In 1935, however, just when this view was gaining general acceptance, Stanley 

 in the United States made the startling announcement that he had succeeded in 

 crystallizing the tobacco mosaic virus. From subsequent work carried out in the 

 United States and in Great Britain (see Stanley 1938a, b, c, d, 1941, Bawden 

 1943), it became clear that this virus, which produces a disease in the Turkish 

 tobacco plant, apparently consists of large nucleoprotein molecules, about ten 

 times as long as they are wide, having a molecular weight of about 17 million, 

 and capable of fitting together lengthwise to form needle-shaped crystals 20-30 jli 

 long. The high molecular weight of this substance — greater than even the largest 

 hsemocyanin molecules — its ability to produce disease in a quantity as small as 

 one thousand-millionth of a gram, combined with the almost perfect parallelism 

 between the amount of protein estimated chemically and the virus activity estimated 

 biologically, set it apart from any other known protein. It was further shown 

 that variants of this virus had a slightly different chemical structure from the 

 parent virus, and that chemical alterations brought about in the molecule by 

 laboratory methods led to changes in its pathogenic properties. 



The shock produced by Stanley's discovery was somewhat mitigated when it 

 was pointed out (Takahashi and Rawlins 1935, Bawden et al. 1936, Bernal and 

 Fankuchen 1937) that the crystals were not true crystals, but two-dimensional 

 liquid crystals formed as the result of linear aggregation of smaller thread-like 

 bodies showing anisotropy of flow. Under these conditions it did not appear 

 necessary to abandon the conception of viruses as living micro-organisms merely 

 because some of them could arrange themselves in orderly rows (Andrewes 1938). 

 When it was shown, however, by Bawden and Pirie (1938) that the bushy stunt 

 virus of tomatoes formed typical three-dimensional dodecahedral crystals, appar- 

 ently composed of spherical isotropic particles, this way of escape from the dilemma 

 proposed by Stanley looked less promising. 



How are we to reconcile the complex bacteria-like structure and behaviour of 

 the larger animal viruses with the apparently pure nucleoprotein molecules of 

 the plant viruses ? The explanation favoured by many workers (Green 1935, 



