310 H. K. SCHACHMAN AND R. C. WILLIAMS 



for information about the tliickness. O'Konski and Haltner (1956), and 

 Norman and Field (1957), using electric birefringence, obtained a length of 

 about 3400 A, and Rowen and Ginoza (1956) and Boedtker and Simmons 

 (1958) obtained 3300 and 3500 A, respectively, from flow birefringence 

 measurements. It should be noted that rotational diffusion measurements 

 are particularly suited for studies of the length of elongated particles because 

 the rotational diffusion coefl&cient varies inversely with the third power of 

 the length. Thus both flow and electric birefringence techniques have given 

 valuable information about the inhomogeneity of the particles with regard 

 to length. Axial ratios and the dimensions of TMV have also been obtained 

 by a combination' of sedimentation and diffusion data. These results are not 

 given here, however, because of the wide disparity among the reported 

 diffusion coefiicients. 



Evidence as to the shape of TMV was also derived from measurements of 

 the angular dependence of the intensity of light scattered by a solution of 

 the virus. The dissymmetry of the scattering envelope showed clearly that 

 the particles were rodlike in solution and not coils or spheres. The earlier 

 results of Oster et al. (1947) gave 2800A for the length, while the more recent 

 investigations (Boedtker and Simmons, 1958) yielded 3200A. 



Despite the insensitivity of the sedimentation coefiicient to length for 

 long rodlike particles, considerable information as to the homogeneity of 

 TMV can be derived from ultracentrifugal studies. For some years there 

 appeared to be a paradox between the results of the workers employuig the 

 ultracentrifuge, on the one hand, and those investigating TMV with the 

 electron microscope, on the other. Whereas the latter workers invariably 

 observed many small particles in their micrographs, the ultracentrifugal 

 studies frequently yielded sharp, symmetrical boundaries with no obvious 

 evidence for the presence of the smaller particles. It should be noted here 

 that the optical system in the ultracentrifuge gives weight fractions, whereas 

 electron micrographs are almost invariably interpreted in terms of number 

 fractions. When tliis difference is taken into accomit some of the discrepancy 

 disappears immediately because a particle having one-third the most 

 common length (3000 A) makes only one-third the contribution to the ultra- 

 centrifuge pattern that the same particle makes to the histogram prepared 

 from electron micrographs. Further evidence aimed at clarifying this ap- 

 parent discrepancy between the results of the two techniques was provided 

 by the ultracentrifugal studies of Schachman (1951b). By exploiting a boun- 

 dary anomaly, the Johnston-Ogston effect (Johnston and Ogston, 1946), 

 Schachman found that particles about two-thirds as long as the most common 

 particle were readily detectable even if these smaller particles were present 

 to the extent of only 2 % of the total weight. By inference this work sup- 

 ported the conclusions of Williams and Steere (1951) to the effect that most 



