44 ON SOME ASPECTS OF THE 



into the body of a specific microbe (bacillus anthracis), 

 and thus laid the foundations of the now extensive science 

 of bacteriology in its application to the science of medicine. 

 The convincing nature of every such crucial experiment 

 is dependent upon its freedom from sources of error, 

 suspected or imperfectly known at the time ; in the 

 carrying out of such experiments the scientific man 

 labours with a special sense of his intellectual and moral 

 responsibility, and the fear of error thus necessitates that 

 the same experiment should be repeated over and over 

 again. In the example I have selected there are various 

 considerations which make this responsibility obvious to 

 the public at large, but a similar responsibility attaches 

 to every experiment carried out for the furtherance of 

 scientific knowledge, necessitating its frequent repetition 

 under similar and under changed conditions. Nor does 

 the responsible duty of the scientific man end here, for 

 the fear of leading others astray is ever pressing upon 

 him and demands that his description of his work shall be 

 in language stripped of all ambiguity, language which shall 

 set forth the experimental procedure and the results in their 

 naked truth. The character of scientific work, with its 

 expenditure of time, labour, and even wealth upon what the 

 world regards as trivial detail, thus requires from him who 

 employs the scientific method ungrudging service ; thus 

 only can he aid in constructing the slowly rising temple of 

 science. ' The philosophy I recommend,' says Robert 

 Boyle,' is much more difficult, laborious, and expensive than 

 that of Aristotle and the Schools/ In dignified language 

 Boyle shows the world and all his carping critics that his 

 own spirit was one prepared at all hazards to undertake 

 its pursuit. ' I am content/ says he, ' provided experimental 

 knowledge be really promoted to contribute to the meanest 

 way to advance it, and had rather be an under-builder 



