312 MUTATION AND PLANT BREEDING 



magnitude than improvements by other modifications and, secondly, 

 chemical purification can only remove what is already present and 

 now approaches complete recovery. With culture development 

 responsible for yield improvement from the 1 to 10 gamma/ml to 

 the 1,000 gamma /ml range, further increase is expected. 



Practically no experimental methods and results are published 

 from industrial laboratories on this topic, since these are trade 

 secrets (59). A complete and detailed account of strain selection for 

 increased penicillin yield has been published (8). The absolute yields 

 are, however, below present industrial production levels. References 

 to published papers concerning strain selection of commercially 

 important antibiotics have been published (2); other papers appear 

 in symposia (14, 41) and review volumes (62). 



No published work clearly separates the effects of culture selec- 

 tion from changes in media, methods of growing inocula, and condi- 

 tions of fermentation on yield of commercially important antibiotics 

 during industrial production. An interesting account of a strain selec- 

 tion program is given by Kashii, et al. (67) using successive mutagenic 

 treatments, isolation, and application of statistics. 



When a decision is made to produce pilot plant quantities of a 

 new antibiotic for pharmacological and clinical tests and determina- 

 tion of structure, a program of strain selection and media develop- 

 ment is usually begun on a laboratory scale. This program then 

 expands to include conditions of inoculum development and fer- 

 mentor tank operation in the pilot plant. Strain selection programs 

 in the laboratory follow a sequence of steps, regardless of the specific 

 organism and antibiotic. These steps are routine and have been in 

 use throughout the period of industrial antibiotic production. These 

 will be described with some speculations on different approaches 

 suggested by recent genetic findings. 



A word of caution concerning measuring quantitative differences 

 in antibiotic yield needs introduction here. The only proof of the 

 worth of an antibiotic-producing culture is the yield obtained under 

 actual production conditions. Since strains with an improved yield 

 from a production level culture are rare, a less expensive system than 

 tank fermentations returning results rapidly on a large scale is neces- 

 sary. The "shake flask" system is commonly used. The antibiotic 

 yield of various isolates is determined in a small volume of medium 



