THE PHYSICAL PROPERTIES OF INFECTIVE PARTICLES 305 



this effect (see review by Schachman (1959) for details). Seldom has this 

 been done for any protein and apparently never for TMV. It is tempting to 

 attribute the spread in values to variations among the preparations, but 

 doing this tends to obscure the limitations in the precision of the measure- 

 ments themselves. To be sure, preparations frequently contain a fraction 

 of the virus in aggregated form, as dimers or higher aggregates, but this 

 circumstance need not interfere with the accurate determination of the 

 sedimentation coefficient of the bulk of the material under investigation. 

 Similarly the presence of partially degraded material in the preparation need 

 not complicate the accurate determination of the sedimentation coefficient 

 of the principal component and its extrapolation to infinite dilution. As 

 indicated in Section II, A, 5, the sedimentation coefficient should be ac- 

 companied by a diffusion coefficient if molecular weights are to be evaluated. 

 On theoretical grounds this is the procedure of choice, since no conceptions 

 about the shape or hydration need be invoked. Unfortunately, the diffusion 

 data for TMV camiot be considered rehable. Among the various determina- 

 tions, the values, 0.53 X 10-' cm.7sec. (Lauffer, 1944a), 0.45 X 10-' 

 (Schramm and Bergold, 1947), and 0.4 X 10"' (Watanabe and Kawade, 

 1953) are probably the most dependable. None of them, however, can be 

 accepted if present standards are applied. Consequently the molecular weight 

 calculated from these data must be considered as only approximate. It is 

 ironic to note that the partial specific volume, which usually is the least 

 precisely determined among the parameters required for molecular weight 

 calculations, is known with greater accuracy than either the sedimentation 

 or diffusion coefficient. The value 0.73 cc./gm. has been obtained by different 

 workers (Bawden and Pirie, 1937; Lauffer, 1944a; Schramm and Bergold, 

 1947). From the data enumerated above we can calculate that the molecular 

 weight hes between 31 X 10« and 45 X 10«. 



In view of the foregoing remarks, and especially because of the lack of 

 critical diffusion studies, none of the values mentioned above should be 

 considered as confirmatory of molecular weights obtained in other ways. 

 Advances in knowledge of the structure of TMV have been extremely rapid 

 in the past few years, but further progress is dependent in part upon accurate 

 information of the molecular weight. Although the means are now available, 

 the sedimentation-diffusion method has yet to be exploited to its fullest 

 capacity. 



The molecular weight of TMV has frequently been computed by combining 

 the sedimentation coefficient with the intrinsic viscosity. This calculation 

 involves assumptions about the rigidity and hydration of the particles and 

 the absence of permeation of solvent through the particles during their 

 movement through the liquid. However, it can be shown theoretically that, 

 if both the viscosity and the sedimentation velocity determinations are 



