RELATIONS TO OTHER SCIENCES 461 



own science, should make themselves first acquainted with the 

 general principles of the method itself. The non-observance of 

 this precaution has led in more than one instance not only to fal- 

 lacious scientific conclusions but to deplorable practical results. 



1 will illustrate this by one example only. 



In no chapter within the territory of our own specialties has 

 the statistical method of late years been more frequently used than 

 in that of cancer of the larynx. As a matter of fact, the usefulness 

 or otherwise of the individual operations now practiced for the 

 cure of that terrible disease is judged by most surgeons exclusively 

 on the basis of statistics recording the results of various forms of 

 operation. Unfortunately, however, a good many of those who 

 have compiled such statistics have done so in a most empirical 

 manner. They have simply registered under one and the same 

 heading all operations of one and the same type ever performed 

 without taking into consideration such indispensable distinctions 

 as: 



1. The period of our knowledge at which each of these opera- 

 tions was performed. 



2. The individual and enormously different conditions present 

 in each of the cases which were subjected to one and the same 

 operation. 



3. The progress of the technique of these operations as they 

 gradually developed. 



The outcome of this, as will be clear to everybody who has paid 

 any attention to the principle of statistics, has naturally been 

 lamentable. Most valuable forms of operation, such as thyrotomy, 

 have been and unfortunately still are persistently discredited, 

 because some compilers of these statistics will not or cannot see 

 that a thyrotomy performed, say, in 1870, was a thing as different 

 as heaven and earth from a thyrotomy performed in 1904 under 

 altogether different conditions of diagnosis and technique. They 

 accordingly put together all thyrotomies ever performed, without 

 taking these all-important differences into consideration, and 

 calmly proceed to register the net result. The natural outcome of 

 such directly misleading statistics has been that the true value 

 of thyrotomy in suitable cases has not nearly universally enough 

 been recognized at the present moment, and those who have prac- 

 ticed it with excellent results in really suitable cases during the 

 last fifteen years have even at this hour of the day to carry on an 

 uphill fight against those who put their faith blindly in the un- 

 satisfactory sort of statistics just described. The hope may there- 

 fore be justly reiterated on this occasion that every medical man 

 who wishes to approach a medical question from the statistical 

 point of view should make himself thoroughly acquainted with 



