COMPLETION OF RESEARCHES. 397 



sometimes put aside in an incomplete state by the investi- 

 gator for the purpose of pursuing new ones, which appear 

 more important or more attractive. 



A complete research should be exhaustive, because it 

 often happens that if a research is not thorough, instances 

 of an important or exceptional kind remain undiscovered, 

 essential similarities or differences continue unnoticed, 

 and the true explanations may remain unattained and 

 a wrong one be assumed. Some scientific men, having 

 made an experiment and observed its result, or observed 

 some singular natural phenomenon, stop short, and fail to 

 examine it further. Others make a partial investigation 

 only, either making an insufficient variety of experiments, 

 and thereby failing to exclude interferences; or an insuffi- 

 cient number, and thus missing extreme and exceptional 

 instances. 



Every part of an investigation, and even each single 

 experiment, may however be viewed as a smaller research, 

 complete . as far as it goes, because we can compare and 

 classify results, and draw analogies and inferences from 

 them as we proceed, and also infer from each result and 

 class of results the full amount of conclusion warranted by 

 the evidence. Viewed from the opposite aspect, the most 

 extensive research is never complete, and never can be 

 until it extends to the utmost bounds of knowledge. 



