196 THE PATH OF SCIENCE 



lead thought today have been educated as engineers. Which 

 school of thought is right? The ans^ver to this depends on 

 what we want to do. 



We can plan for the future and then we can carry out our 

 plans provided that we remember the limitations of planning. 

 We can only plan things that we can control, and our plans 

 will be carried out only so long as our control is effective. We 

 can plan production in a factory because we can control it. 

 If the production is falling below our needs, we can increase 

 it; if it exceeds them, we can diminish it. To plan, \ve need 

 two things: first, the kno'^vledge of the processes that we are 

 attempting to control; second, the physical power to control 

 those processes. It is when we extend our planning from the 

 things that we know to the fields where our knowledge is 

 weak and from the things that we can control to those that 

 are in their nature uncontrollable that our planning fails. 



When these principles are applied to the planning of scien- 

 tific research, we find that the kinds of research that can be 

 planned best are those which are least fundamental. Pro- 

 duction can always be planned. The last stages of develop- 

 ment can be planned with considerable certainty. When a 

 new chemical has been made in the laboratory and the yields 

 have been tested, a pilot plant must be built. The building 

 of this pilot plant and even the time which it will take to test 

 the processes on a moderate scale can be foreseen, and so in 

 chemical factories pilot plant operation is usually carried out 

 not as a research experiment but as a co-operative effort in- 

 volving both the research men who originated the process 

 and the production men who will operate it. Not infre- 

 quently the whole is under the direction of a chemical engi- 

 neering group who specialize in pilot plant operations. 



When more basic research is considered, planning neces- 

 sarily becomes less certain. If ^ve have made a new chemical 

 in the laboratory, we know that we can make it in a pilot 

 plant in spite of the fact that new problems may arise. But if 

 the chemical has never been made or even if it has been made 

 but the yields are unsatisfactory, we know less certainly ho^v 



