THE ART OF SCIENTIFIC INVESTIGATION 



scientific method omitting any reference to chance or empiricism 

 in discovery. 



Perhaps the most striking examples of empirical discoveries are 

 to be found in chemotherapy where nearly all the great discoveries 

 have been made by following a false hypothesis or a so-called 

 chance observation. Elsewhere in this book are described the 

 circumstances in which were discovered the therapeutic effects of 

 quinine, salvarsan, sulphanilamide, diamidine, paraminobenzoic 

 acid and penicillin. Subsequent rational research in each case 

 provided only relatively small improvements. These facts are the 

 more amazing when one thinks of the colossal amount of rational 

 research that has been carried out in chemotherapy. 



The research worker should take advantage of this knowledge 

 of the importance of chance in discovery and not pass over it 

 as an oddity or, worse, as something detracting from the credit 

 due to the discoverer and therefore not to be dwelt upon. 

 Although we cannot deliberately evoke that will-o'-the-wisp, 

 chance, we can be on the alert for it, prepare ourselves to 

 recognise it and profit by it when it comes. Merely realising the 

 importance of chance may be of some help to the beginner. We 

 need to train our powers of observation, to cultivate that attitude 

 of mind of being constantly on the look-out for the unexpected 

 and make a habit of examining every clue that chance presents. 

 Discoveries are made by giving attention to the slightest clue. 

 That aspect of the scientist's mind which demands convincing 

 evidence should be reserved for the proof stage of the investiga- 

 tion. In research, an attitude of mind is required for discovery 

 which is different from that required for proof, for discovery and 

 proof are distinct processes. We should not be so obsessed with 

 our hypothesis that we miss or neglect anything not directly 

 bearing on it. With this in mind, Bernard insisted that, although 

 hypotheses are essential in the planning of an experiment, once 

 the experiment is commenced the observer should forget his 

 hypothesis. People who are too fond of their hypotheses, he said, 

 are not well fitted for making discoveries. The anecdote (related 

 in Chapter Eight) about Bernard's work starting from the 

 observation that the rabbits passed clear urine, provides a beauti- 

 ful example of discovery involving chance, observation and a 

 prepared mind. 



32 



