344 C. E. SCHWERDT 



sedimentation rate yields an estimate of the virus particle density. Sucrose, 

 bovine serum albumin, and DgO have been used to increase the density of the 

 suspending medium. The results obtained with DgO are the easiest to inter- 

 pret since mixtures of DgO and HgO act as a one-component solvent and the 

 particle density measured in such a system is the dry weight density essential 

 for the computation of the virus particle mass and numbers per unit volume 

 of suspension as described above. 



All of the above methods of measuring particle diameter and density have 

 their inherent sources of error. Assumptions made in the calibration of 

 average pore diameter of gradacol membranes and adsorption of virus to 

 pore walls give cause for uncertainty in ultrafiltration studies. The flattening 

 of virus particles as well as possible shrinkage during dehydration in prepara- 

 tion for electron microscopy make this method of measuring virus particle 

 diameter equally uncertain. The estimations of particle dry weight density 

 and diameter by sedimentation analysis in mixtures of DgO and H2O may 

 easily be in error because of the long and hazardous extrapolation necessary 

 to estimate the solvent density at which zero sedimentation occurs (Schach- 

 man, 1957). The extrapolation is long because the density of DgO is approxi- 

 mately 1.1 while the dry weight density of viruses may be as great as 1.35 (or 

 greater), depending upon their nucleic acid content. Since the volume of a 

 spherical particle varies with the cube of the radius, even small errors in the 

 measurement of particle diameter greatly magnify errors in the estimation of 

 particle mass. Sharp (1953) points out that particle diameters obtained from 

 sedimentation velocity, ultrafiltration, and electron microscope data on a 

 given virus which agree to better than ± 5 % are probably accidental and 

 that even this conservative estimate of the error represents a 30 % difference 

 in particle mass. 



Recognition of these difficulties has resulted in the replacement of this 

 indirect method of estimating particle concentration by the direct counting 

 procedures made possible by electron microscopy. Before the latter techniques 

 had been developed, however, vaccinia (Smadel et al., 1939), Shope papilloma 

 (Bryan and Beard, 1940a), and influenza (Friedewald and Pickels, 1944) 

 virus particle counts had been calculated from estimates of particle mass. 



2. Based on Estimate of Number of Hemagglutinating Particles 



Attempts have been made to estimate the number of hemagglutinating 

 particles of influenza and Newcastle disease viruses on the assumption that a 

 single particle may cause the agglutination of two red blood cells. The 

 concentration of red cell dimers which will sediment more rapidly than single 

 cells can be estimated by simple colorimetric means. The procedure is carried 

 out as follows: A small amomit of virus is mixed with a suspension of red 

 blood cells of known concentration. The concentration of virus must be such 



