DIFFICULTIES 



W. B. Cannon^^ comments on the false deduction once made 

 that adrenahne does not play a part in controlUng the sugar 

 level in the blood by calling forth sugar from the liver, on the 

 ground that the blood-sugar level is maintained after removal 

 of the adrenal medulla. The fact is that there are other methods 

 of mobilising sugar reserves from the liver but none are so 

 effective as adrenaline. Shivering by itself can prevent body 

 temperatures from falUng, but that does not prove that other 

 processes cannot play a part. A variant of this " fallacy of a 

 single cause " has been described by Winslow.^*"^ When a 

 combination of two factors causes something, and one is 

 universally present, it is usually rashly concluded that the other 

 is the sole causal factor. In the nineteenth century it was 

 believed that insanitary conditions in themselves caused enteric 

 fever. The causal microbes were then universally present and 

 the incidence of the disease was determined by presence or 

 absence of sanitation. The cause of a disease is complex, 

 consisting of a combination of causal microbe, the conditions 

 necessary for its conveyance from one host to the next and 

 factors affecting the susceptibility of the host. Any happening is 

 the result of a complex of causal factors, one of which we usually 

 single out as the cause owing to its not being commonly present 

 as are the other circumstances. 



Wrong conclusions about the incidence of some condition 

 in a population are sometimes drawn through basing the observa- 

 tions on a section of the population which is not representative 

 of the whole. For example, certain figures were generally 

 accepted and printed in text-books as an index of the proportion 

 of children at different ages that gave a negative reaction to 

 the Schick test for immunity to diphtheria. Many years later 

 these figures were found to be true only for children of the 

 poorer classes attending public hospitals in the city. The figures 

 for other sections of the population were very different. When 

 I went to the U.S.A. in 1938, scarcely anyone I met could say 

 a good word for President Roosevelt, but Dr. Gallup's method 

 of sampling public opinion showed that more than fifty per cent 

 supported him. There is a great temptation to generalise on 

 one's own observations or experience, although often it is not 

 based on a sample that is truly random or sufficiently large to 



"7 



