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



general order of 30 A can be reliably observed, it is only recently that tliis level 

 of detail has been considerably reduced. Hall (1956) has developed a method 

 involving preshadowed replication (Williams and Wyckoff, 1945), wherein a 

 surface of freshly cleaved mica is used as a structureless substrate surface 

 upon which the specimen objects are dispersed. Shadowmg is done with 

 platinum, over which is deposited a film of silicon monoxide or carbon. Sub- 

 sequently the shadowing film and its overlay are stripped from the mica with 

 the aid of a supporting collodion film. In principle the method is quite old; 

 the excellence of HaU's results seems to reside in the use of mica and in the 

 application of a film of great stability (such as silicon monoxide) over the 

 shadowing one. 



The preservation of the details of virus structure has always been a 

 problem in electron microscopy, brought about by the necessity of observing 

 the virus particles after they have dried from their aqueous suspending 

 medium. Two problems are involved here: one is the elimination of non- 

 volatile salts that are frequently fomid in a suspending medium, while the 

 other is the prevention of distortion of the virus particles as they dry. Viruses 

 are frequently suspended in solutions of appreciable ionic strength, and if the 

 ions are those of nonvolatile salts, such as NaCl, a virus suspension when 

 dry will exhibit far more salt than virus. A way out of this difficulty has been 

 found by the use of suspending media whose ionic constituents are com- 

 pletely volatile (Backus and Williams, 1950). At the present time it has been 

 fomid that ammonium acetate, bicarbonate, and benzoate are particularly 

 useful salts to use m those cases where pure water is not an adequate 

 suspending medium. Another method that has proved effective m the elimina- 

 tion of nonvolatile salts in a dried virus preparation is one in which a suspen- 

 sion is allowed to dry on a film of collodion that is in contact with a block of 

 semisolid agar (Kellenberger and Kellenberger, 1955). The effect of the agar 

 is to imbibe the salt ions along with the water molecules, leaving the larger 

 virus particles on the surface of the coUodion. 



It is readily seen that when a particle as small and as nonrigid as a virus 

 dries out of a water suspension it is subject to considerable pressure brought 

 about by surface tension forces. The effect of this pressure is to flatten the 

 particle and to distort its structure (Fig. 7). The alleviation of this artifact 

 has taken two directions, by Anderson (1951), and by WiUiams (1953a). In 

 the method developed by Anderson the virus particles are first seen exposed to 

 osmic acid and are then brought through various miscible solvents into carbon 

 dioxide at a pressure and temperature such that it is liquid. When the tempera- 

 ture is raised it passes through its critical point, and the virus particles now 

 find themselves in a gaseous environment. After the COg gas is allowed to 

 escape the virus particles are ready for observation, having been dried without 

 exposure to the forces of aqueous surface tension. The method developed 



