FILTRATION AND ULTRAFILTRATION 



951 



It is important to realize that the mere passage of an organism through a filter 

 candle does not justify its inclusion in the group of filtrable viruses. Even under con- 

 ditions of careful experimentation, small organisms, particularly slender flexible and 

 motile organisms such as spirochsetes, frequently appear in small numbers in the fil- 

 trate ; and conversely, the mere failure of an organism to pass through a filter candle 

 does not justify its exclusion from the group of filtrable viruses. Some viruses, for 

 example, such as those of varicella and herpes zoster, have not yet been shown to 

 be filtrable, yet there is little doubt from what is known of their other properties 

 that they should be included in this group. The term " filtrable virus " is one 

 connoting a number of properties, the most important of which have already 

 been defined at the head of this chapter. 



Ultrafiltration.— In recent years ultrafiltration has been introduced. In this 

 process, thin collodion membranes are prej^ared with a given size of pore, the size 

 being determined largely by the concentration of collodion used. In the develop- 

 ment of these filters Elford (1931, 1933) has played a prominent part. Starting 

 from the earlier work of Bechhold, he has been able by the use of appropriate solvent 

 mixtures, and by the careful standardization of his technique, to prepare a series 

 of membranes of very regular and accurately graded porosities by means of which 

 determinations of the size of many of the commoner viruses have been successfully 

 made, and subsequently confirmed by other workers (see also Elford, Grabar, and 

 Ferry 1935, Duclaux and Amat 1938). These filters — Gradocol membranes — 

 approach nearer to the mechanical sieve than do ordinary filter candles ; they 

 appear to be less influenced by the various secondary factors which we have men- 

 tioned, and to be capable, when properly used, of sorting out particles very largely 

 according to their size ; though the influence of the pH of the suspending fluid, 

 and of the electrical charge carried, must still be taken into account. 



In calculating the size of the particle from the average pore diameter through 

 which it just fails to pass, the effect of adsorption has to be considered. This effect 

 is most influential in membranes with very small pores (Table 58). Thus a 

 particle held back by a membrane with an average pore diameter of 30 m/< probably 

 has a diameter of 10-15 m/^, while one held back by a membrane of 1,000 mjx 

 probably has a diameter of 0*75-l-0^ (Elford 1933). This relationship, however, is 

 disputed by Markham, Smith and Lea (1942) (see also Cox and Hyde 1932, Asheshov 

 1933 on ultrafiltration). 



TABLE 58 



Relation of Size of Retained Particle to average Pore Diameter of Gradocol Mem- 

 branes (Elford 1933) 



d — average pore diameter of limiting membrane for optimum filtration conditions. 



Microscopical Examination. — There are difficulties in the microscopical examina- 

 tion of viruses. It has already been pointed out in Chapter 2 that under ordinary 



