5. 



would not stop the practice. It is unnecessary because there is an 

 effective course of action open to us, \vhich is to make as careful an 

 assessment of the situation as possible, noting in detail the areas in 

 which more information is needed. We can then recommend proced- 

 ures that correspond to the conservative, or safe, side of the large 

 band of uncertainty that must surround the predicted behavior of the 

 system we are considering. We may rely upon the results of the 

 recommended studies to furnish the basis for revision of our conclu- 



