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



precise, molecular weights of TMV calculated from these data will not be 

 much in error even if the virus particles are swollen to have a shape markedly 

 different from that characteristically seen in the electron microscope. Even 

 in the worst case the error in the molecular weight might be expected to be 

 no more than 20 %. But if a degree of precision around 1 % is desired, then 

 it is necessary that the assumed model of the particle in solution conform to 

 the real particle. 



The intrinsic viscosity provides a sensitive measure of the degree of 

 aggregation of the virus particles. This was demonstrated by Lauffer (1944a) 

 when he compared, for different preparations, the intrinsic viscosity with 

 the length-distribution curve evaluated from electron micrographs. In this 

 way he was able to provide good experimental evidence for the validity of 

 the Simha equation (Simha, 1940). Because of the tendency of the virus 

 particles to aggregate end-to-end to form dimers and even higher aggre- 

 gates, it is not surprising that the intrinsic viscosities reported in the litera- 

 ture exhibit wide variation. Even for preparations which seem to contain 

 predominantly unaggregated material values between 0,28 (gm,/100 ml.)~^ 

 (Lauffer, 1944a; Schachman and Kauzmann, 1949) and 0,37 (gm,/100 ml.)-^ 

 (Watanabe and Kawade, 1953; Boedtker and Simmons, 1958) have been 

 obtained. These values, in conjunction with the sedimentation coefficients 

 given above, yield 35 X 10^ to 44 X 10^ for the molecular weight. For this 

 calculation a value of ^ = 2.5 X 10^ has been used in Equation 24, The 

 reader may find it profitable to repeat this calculation for particles of 

 different assumed axial ratios. 



Molecular weights for TMV have also been obtained by light scattering. 

 The most recent determination (Boedtker and Simmons, 1958) gave the 

 value, 39.0 ± 1.2 X 10^, which seems to be in excellent agreement with the 

 value obtained ten years earlier by Oster et al. (1947). In comparing these 

 values, however, it is important to note that the newer measurement of the 

 specific refractive increment (dn/dc) gave a value 15 % greater than the 

 older value. It is general custom to correct older results from light-scattering 

 investigations by inserting the more recent value of dn/dc, since the deter- 

 mination of dnjdc is much more accurate now than it used to be. Doing this 

 would cause a decrease of 24 % in the molecular weight reported in the 

 earlier investigation and would destroy the apparent agreement mentioned 

 above. Also the results of Doty and Steiner (1950) would require revision, 

 leading to values below 40 X 10^. In all tliree investigations it was claimed 

 that aggregation of the virus had not occurred to a significant extent. It is 

 clear that these results, like those from sedimentation velocity studies, 

 differ by amounts much larger than the presumed experimental errors. 

 Again it is tempting to attribute this to differences among the samples, but 

 a careful reappraisal of the experimental aspects of light scattering by TMV 



