260 GENETICS AND PLANT BREEDING 



factorial analysis of bacterial heredity lent and lack the specific capsular sub- 

 has, of course, had to wait upon the stances. Some rough variants were 

 recognition of recombinational tech- more or less unstable and would occa- 

 niques, be they sexual or transdue- sionally revert to the parental smooth 

 tional. Unit factors that are subject to form; others appeared to be absolutely 

 mutation, and can be used for genetic stable. Grifhth was impressed by the 

 studies of bacteria, are related to such occasional occurrence of more than one 

 diverse traits as pigmentation, resist- serotype in a single sputum, and specu- 

 ance to antibiotics and to bacterial vi- latcd on the possible interconvcrtibility 

 ruses, biochemical aspects of nutrition of the types. His experimental design 

 and fermentation enzymes, cellular was fundamentally similar to that of 

 morphology, and antigenic specific- earlier attempts at transduction: he 

 ity. . . . prepared heat-killed suspensions of 

 The idea of transduction of heredi- smooth bacteria, and inoculated mice 

 tan- fragments is still difficult to rccon- with a mixture of this vaccine and 

 cile with our well established knowl- some living, rough cells. Neither the 

 edge of highly organized chromosomal killed vaccine nor the rough bacteria, 

 systems of heredity, but the problem separately, would be expected to gen- 

 can now be stated in terms clear crate an active infection in the mice, 

 enough that the synthesis, if not yet and none did in control experiments, 

 certain, can at least be experimentally but in a few inoculations, the mixture 

 sought. ... of the two gave a virulent infection 



from which living smooth pncumo- 



THE "PNEUMOCOCCUS cocci could be isolated. In further ex- 



transformation" periments, he also showed that the 



serotype of the recovered bacteria de- 



A new turn of events was marked pended on the type of bacteria used to 



by Griffith's announcement in 1928 prepare the vaccine, rather than on the 



(the same 3ear as Muller's proof that parentage of the rough cells. He had 



X-rays caused mutations ) of the trans- therefore justified the claim that rough 



formation of serological types in the bacteria, originally of one type, had 



pneumococcus. He had been working been transformed to a different smooth 



for some \ears on the serotypic classi- type. The mouse played the part of a 



fication of pneumococci, which was im- selective agent in obtaining a specific 



portant for therapy as well as for diag- genotype, namely a new virulent form, 



nosis, since antiserum was almost the To explain this transformation, Grif- 



only specific treatment known at that fith did not use genetic language, but 



time. Tlie most prominent antigen of adopted what amounts to transduction 



virulent pneumococci consists of a hypothesis, namely that the substances 



polvsaccharide capsule that envelopes responsible for the formation of the 



each cell, and is readily detected by an specific capsules had been transferred 



apparent "Ouellung" (swelling) reac- via the vaccines to the deficient, rough 



tion in the presence of specific anti- cells. 



serum. . . . The presence of the cap- Fortunately, Griffith's observations 



sules conditions not only the high were saved from limbo by their prompt 



virulence of these organisms, but also trial in several other laboratories where 



the appearance of their colonies on they could be confirmed and extended, 



agar media, whence these are called The problem was taken up in the 



"smooth" tvpes. "Rough" variants are laboratories of the Rockefeller Insti- 



also known- these are relatively aviru- tute by a group of workers under the 



