Growth Measurements 165 



bility very nearly equal to that obtainable in 

 making dry weight determinations and has the 

 great advantage that it can be practiced rapidly 

 and without removing the cultures from their 

 flasks. Repeated measurements can be made on 

 each growing culture at any desired interval, 

 growth curves can be set up, and subcultures can 

 be made from those roots which show, over an 

 extended period, any desired type of behavior. 



Replications. As we have already seen, root 

 and callus cultures, since they are biological mate- 

 rials, show a considerable degree of variability. 

 In the series presented in the last table, the lengths 

 at the end of a week in the control nutrient varied 

 between 37 and 178 mm., with an average of 131 

 mm. This is fairly typical of root cultures. It is 

 obvious that materials of this degree of variability 

 cannot be treated as individuals but must be 

 treated statistically. This being the case, it is 

 important to decide first how many replications 

 are necessary in order that significant results can 

 be obtained and, second, how wide a difference 

 must a given experimental variable produce in a 

 series of the chosen number of cultures before that 

 effect can be considered as being due to this vari- 

 able and not to chance. Decision on the first point 

 involves a compromise between the ideal of an 

 infinite number on which to base averages and the 



