164 BULLETIN OF THE 



other is the faculty of eliminating accidental conditions from 

 those which are essential : and farther, the characteristic of perse- 

 verance is indispensable. 



The fields of scientific labor may be divided into two classes, 

 viz. : those which relate to the empirical observation of facts, and 

 those which refer to the systematic series of investigations rela- 

 tive to the law or cause of special phenomena. As illustrations 

 of the first class are the facts of the phenomena of the physics 

 of the globe, those of ordinary meteorology and natural history; 

 while, as examples of the second, we have the phenomena of 

 chemistry, physics, and astronomy. 



The remarks which I have previously made relate principally 

 to the former. In order to elucidate the method of investigation 

 in the latter case, I will suppose the existence of a new pheno- 

 menon which is unconnected with any of the present generaliza- 

 tions of science, but of which it is desired to discover the law or 

 the facts with which it is associated. Such facts standing alone 

 form no part of science ; they are usually discovered in the course 

 of investigations, and are of great importance in pointing out 

 fields of new research which promise an abundant harvest. 



The first step in the investigation is to reproduce the phenome- 

 non ; the next, is to form in the mind a provisional hypothesis 

 as to its cause, and in the choice of this we are governed by 

 analogy. For example, if it appears to resemble some of the phe- 

 nomena of electricity, we assume that it is produced by electri- 

 city ; we next endeavor to ascertain by what known action of 

 electricity such an effect could possibly be produced ; for this 

 purpose we invent an hypothesis, or imagine some peculiar actioa 

 of electricity sufQcient to produce the effect in question : we then 

 say to ourselves, if this be true, it will logically follow that a, 

 specific result will follow if we make a certain experiment : the 

 experiment is devised and tried, but no positive result is obtained. 

 In order to this negative result, the logical deductions must have 

 been in error, or the experiment must have been defective, or the 

 hypothesis itself erroneous. 



We examine each of the two former steps, and finding nothing- 

 amiss in them, we conclude that the hypothesis was not true : 

 another hypothesis is then invented, another deduction inferred, 

 and another experiment made ; still no result is obtained. At 

 this stage of the research the inexperienced investigator is prone 



