554 THE ROYAL SOCIETY OF CANADA 



the insertion which he purposes to make should be at a point and of a 

 depth that might involve a vital part if the organs were in fact reversed, 

 he would not be justified in assuming without further test that the 

 organs are in the usual position. 



In the first case a probability of 200,000:1, or 20,000:1, or even 

 of 8,000:1, is considered ample to justify action without further test, 

 because these are all formal probabilities, or the product of a conscious, 

 competent operator of correct processes; in the second, a probability 

 of 1,000,000:1 is not sufiicient because this is only pure probability; 

 for nature is not a conscious, intelligent operator, whose occasional 

 failures can be attributed to inadvertent or incidental error; and 

 therefore, if even but one failure is registered against one of her natural 

 processes, such as that for placing the human organs in the usual 

 position, her success in achieving this usual result is thereby reduced 

 to the blind action of many favorable, against a few unfavorable, 

 chances. 



The question may suggest itself, "Would not the surgeon also 

 make further test in such a case, if the probability of 1,000,000:1 were 

 formal, instead of pure ?" Nature being never an intelligent operator 

 but having only automatic processes which are either correct or in- 

 correct {i.e., correct, where no failures have been registered against 

 the process, and incorrect where a failure or failures have been regis- 

 tered against it), such a case of formal probability could never arise 

 in this particular connection. We have also no way of deciding this 

 question on the basis of actual experience, because no conscious, in- 

 telligent operator has ever been known to have raised his application 

 of correct processes in any branch of science to an average of correct- 

 ness as high as 1,000,000:1. On these two grounds, the one particular 

 and the other general, the suggested question is to be classed as in 

 itself fictitious. A negative answer to it is, nevertheless, fully justi- 

 fied by precedent and experience. There is not the least reason to 

 suppose that a surgeon would hesitate to act upon a formal probability 

 of 1,000,000:1, or that any one would feel that he should make further 

 test of it; for much lower formal probabilities than this are constantly 

 acted upon without further test in equally critical connections. Thus, 

 in the case cited in Part I of this paper (p. 143), the formal probability 

 that a druggist will compound a prescription correctly is not nearly 

 1,000,000:1, and an error in this respect might be fatal; yet even an 

 analytical chemist, who is well able to test the actual contents of any 

 such compound supplied to him by a druggist, will take the remedy 

 without making this test unless there are manifest grounds for sus- 

 pecting an error by reason of an unusual appearance or odor in the 

 mixture. 



