THE PHYSICAL PROPERTIES OF INFECTIVE PARTICLES 309 



but tlie preparative conditions resulted in aggregation to particles of uniform 

 length; (2) the virus was polydisperse, but the preparative methods pre- 

 served a selection of monodisperse particles at the expense of the shorter 

 ones; (3) the virus in the crude juice was essentially monodisperse. If one is 

 unwilling to accept the third assumption one would have to explain the 

 results of Williams and Steere on the basis of the first assumption, unlikely 

 as it is, since there was no chance of the method having resulted in a selection 

 of monodisperse particles. The second assumption cannot be definitely ruled 

 out as an explanation of Hall's results unless it can be shown that the 

 purification method retains most of the viral starting material. However, as 

 is well known, a centrifuge discriminates poorly between rod-shaped par- 

 ticles of differing length when a length of about 1200 A, for TMV, is reached 

 or exceeded. It appears that the most likely explanation of the agreement 

 between the results of Hall and of Williams and Steere is that their experi- 

 ments sampled without bias the contents of the suspensions and, inasmuch 

 as the preparative conditions were quite different, that the suspensions were 

 representative of the material in the crude juice. Furthermore, since it is 

 hard to imagine that the extraction process alone would introduce mii- 

 formity of length where there was none before, the most reasonable supposi- 

 tion to be made at this time is that the particles of TMV in vivo are 

 essentially monodisperse and are 3000 A long. 



Other methods also have given information about the length of the virus 

 particles. These methods, such as viscometry and streaming birefrmgence, 

 are not direct like electron microscopy and some ambiguity is involved in 

 the interpretation of the data. Basic to the determination of the length of 

 TMV particles by the hydrodynamic techniques is the view that the particles 

 imbibe little or no water. This contention arises largely from the X-ray 

 diffraction studies of Bernal and Fankuchen (1941), who showed that some 

 of the intramolecular spacings of completely dried virus were the same as 

 those in wet gels of T]\IV. This view gained support from the results of 

 ultracentrifugal studies of TMV in solutions of different densities (Schachman 

 and Lauffer, 1949). If the absence of hydration is taken for granted, the 

 physical chemical data can be interpreted directly in terms of the shape of 

 the particles. For this purpose ellipsoidal models are considered, since most 

 hydrodynamic theories deal with particles of that type rather than with 

 cylinders. Thus perfect agreement with the electron microscopic evidence 

 cannot be expected. 



From the viscosity data enumerated above an axial ratio between 20 and 

 24 is calculated. In conjimction with the thickness, 152A, evaluated from 

 X-ray diffraction, various workers have inferred values between 3000 to 

 3600A for the length of the particles. Length measurements have been 

 obtained directly from rotational diffusion studies without any requirement 



