nelson: screening methods in microbiology 315 



followed by dilution and plating to obtain well separated colonies 

 will usually give pure colonies since the conidia probably contain a 

 single haploid genome. Mutagenic treatments applied to germinating 

 conidia and to mycelial growth will give genetically mixed colonies 

 derived from several oenomes. Testing such isolates is fruitless. An 

 intermediate growth step with sporulation and replating is necessary 

 to eliminate this error when the mutagenic treatment is applied to 

 growing cultures. It is convenient to use this step to measure the yield 

 of several classes of easily recovered variants (resistance to antibiotics 

 and phage) as a function of the treatment method. These "muta- 

 genicity indices" may be used to maximalize production and recovery 

 of mutants. While a generalization of effective conditions to muta- 

 tions affecting yield may not be valid, it is possible to eliminate 

 inefficient methods which would prevent the recovery of any isolates 

 with improved yield. 



Small yield improvements may sometimes be obtained by ran- 

 dom selection of colonies plated out from production fermentor 

 tanks. Such isolates may be returned to production fermentor 

 tanks without more than one shake flask test showing yields at or 

 above production levels. This program can often be coupled with 

 continuous changes in operation of fermentor tanks and minor 

 changes in media constitution in a form of evolutionary operations. 

 Unfortunately, a statistical interpretation of results on either a labo- 

 ratory or a production scale is not often possible since projected 

 demands for material for processing and the cost of operating non- 

 productive equipment as "controls" may prevent carrying out valid 

 experiments. 



Random isolation methods are distinguished from other tech- 

 niques by the fact that the culture isolated is not known to be a 

 variant. So many tests of different isolates may have to be made as to 

 make random isolation an inefficient program. It is advisable to set 

 a limit to the number of isolates to be run with various treatments 

 to avoid an unrewarding search. 



Isolation of Variants with Known Physiological Differences 



The possibility of determining the potential yield of isolates 

 without an initial quantitative test in shake flasks has occupied most 

 industrial workers, but no adequate method has yet been reported. 

 A technique commonly used is the "streak plate", on which several 



