THE PHYSICAL PROPERTIES OF INFECTIVE PARTICLES 311 



of the short particles observed in the micrographs were the result of breakage 

 of the longer particles. The ultracentrifugal studies have since been extended, 

 and the results show clearly that the ultracentrifuge, if used in this way, 

 can detect very small amomits of particles shorter than the usual length. 

 Through this type of study and the more conventional boundary spreading 

 analyses a great deal of information regarding homogeneity can be derived 

 from the ultracentrifuge patterns. It is to be deplored that workers are still 

 apparently content with the presentation of an ultracentrifuge pattern 

 which by itself gives no assessment of the homogeneity of the sample. 



There is some direct evidence as to the length of the particles of TMV as 

 they exist within the inclusion crystals commonly found in the hair-cells 

 of infected tobacco plants (Steere, 1957). By the use of a technique that 

 provides a replica of a frozen, lightly sublimed surface that has been cut 

 through a crystal, it has been fomid that the virus particles are arrayed in 

 palisades, and that all the parallel particles in each array are about 3000A 

 long. It is possible, of course, that only the rods of uniform length are able 

 to be formed into the crystals. The observed monodispersity within the 

 crystals does not necessarily imply monodispersity of the particles as they 

 exist tliroughout the cell. Unfortunately, particles of TMV are difficult to 

 discern in sections of plant tissue fixed and embedded in the usual manner 

 for electron microscopy, and consequently nothing is known about their 

 lengths when observed in this way. 



Electron microscopy, combined with biological assay, has given informa- 

 tion about the relation of infectivity to the length of TMV particles. Many 

 experiments have consistently shown that preparations containing a large 

 proportion of rods distinctly shorter than 3000A are less infectious than those 

 that are essentially monodisperse at 3000A. It is known, further, that aggre- 

 gation of short particles to long ones will not enhance uifectivity. The 

 experiments relating infectivity with length are not sufficiently precise, 

 however, to demonstrate whether there is a critical minimal length for 

 uifectivity or whether, as some would prefer to say (Pirie, 1957), the shorter 

 rods have a lower "probability" of being infectious. 



Through the use of the separation cell in the ultracentrifuge, Lauffer 

 (1943a) obtained important information relating the infectivity to the 

 particle size. He showed that the sedimentation coefficient of the infectious 

 agent was similar to that of the principal component in the preparation, and 

 he was able to exclude an hypothesis that infectivity was restricted to 

 particles having a weight, for example, less than one-half the predominant 

 molecular species. 



d. Diameter and Cross-Sectional Shape ; Surface Structure. Both electron 

 microscopy and X-ray analysis are capable of yielding precise information 

 about the diameter of the TMV rods, and less exactly, information about 



