THE ART OF SCIENTIFIC INVESTIGATION 



been able to show any benefit from patulin and why it seemed 

 to be efficacious in the first experiment remains a mystery.^* 



When I saw a demonstration of what is known as the Mules 

 operation for the prevention of blowfly attack in sheep, I realised 

 its significance and my imagination was fired by the great 

 potentialities of Mules' discovery. I put up an experiment involv- 

 ing thousands of sheep and, without waiting for the results, 

 persuaded colleagues working on the blowfly problem to carry 

 out experiments elsewhere. When about a year later, the results 

 became available, the sheep in my trial showed no benefit from 

 the operation. The other trials, and all subsequent ones, showed 

 that the operation conferred a very valuable degree of protec- 

 tion and no satisfactory explanation could be found for the 

 failure of my experiment. It was fortunate that I had enough 

 confidence in my judgment to prevail upon my colleagues to put 

 up trials in other parts of the country, for if I had been more 

 cautious and awaited my results they would probably have 

 retarded, the adoption of the operation for many years. 



Several large-scale experiments in the U.S.A. proved that 

 immunisation greatly reduced the incidence of influenza in 1 943 

 and again in 1945, yet in 1947 the same type of vaccine failed. 

 Subsequently it was found that this failure was due to the 1947 

 strain of virus being different from those current in earher years 

 and used in making the vaccine. 



It is not at all rare for scientists in different parts of the world 

 to obtain contradictory results with similar biological material. 

 Sometimes these can be traced to unsuspected factors, for 

 instance, a great difference in the reactions of guinea-pigs to diph- 

 theria toxin was traced to a difference in the diets of the animals. 

 In other instances it has not been possible to discover the 

 cause of the disagreement despite a thorough investigation. In 

 Dr. Monroe Eaton's laboratory in the United States influenza 

 virus can be made to spread from one mouse to another, but in 

 Dr. C. H. Andrewes' laboratory in England this cannot be 

 brought about, even though the same strains of mice and virus, 

 the same cages and an exactly similar technique are used. 



We must remember that, especially in biology, experimental 

 results are, strictly speaking, only valid for the precise conditions 

 under which the experiments were conducted. Some caution is 



24 



