21 



occurring in a population consuming both mussels and fried 

 fish, with the result that a causal association is set up between 

 the fried fish and the disease. As a matter of fact the first 

 population — that is, the one in which mussels were regarded 

 as the cause of the disease — was also a fried-fish consuming one. 

 We can hardly doubt that much the same conditions obtained 

 in both populations and that there were various ways in 

 which enteric fever might have been communicated, but that, 

 in each case, the Health Officers took one particular aspect 

 of the whole problem. In the first group they were influenced 

 by the Bulstrode tradition, but in the other a spirit of 

 scepticism with regard to accepted methods was allowed to 

 gain force. 



It is quite fair to say that the evidence which imphcates 

 mussels is " coincidence." A man eats mussels and, a week 

 to a fortnight later, he shows that he is suffering from enteric 

 fever. The typhoid germs require a week to a fortnight to 

 " incubate " in the man's body and cause the symptoms of the 

 disease. That may be " mere coincidence," but all scientific 

 proofs are based on just such associations, coincidences in regard 

 to events that happen simultaneously or after a certain period 

 of time. The probability of a causal connection between two 

 such events is small, and it has to be strengthened by the 

 establishment of a series of other coincidences such as, for 

 instance, those that were observable in the cases of the out- 

 breaks following the mayoral banquets at Winchester and 

 Southampton. The strength of the evidence in the latter 

 cases was due to the number of coincidences, and its weakness, 

 in the cases A to E^ on pp. 18-19, is due to their paucity. 

 A certain shop sells mussels which are, presumably, all taken 

 from the same contaminated shellfish bed, and a man buys 

 and eats these mussels and then takes enteric fever a week 

 to a fortnight later. But we ought not to overlook the other 

 coincidences, which seem to me to have just the same value 



