HOW TO OBSERVE AN EXPERIMENT. 317 



During the progress of an experiment, as our perceptive 

 faculties are very limited in delicacy, all the senses of the 

 investigator must be on the alert in order to detect un- 

 expected effects, as well as to observe those which are 

 anticipated, or for which the experiment was purposely 

 made, otherwise small and disguised effects will often pass 

 unnoticed, and we must not forget that small effects^ if 

 they are unusual ones, are often the signs of great general 

 phenomena. When Faraday first discovered magneto-electric 

 action, the effects he obtained were so small that he could 

 scarcely perceive them. 



During the experiment also, all the results, and all the 

 alterations made, and their effects, should be faithfully 

 and truly recorded ; and we must have no preconceived 

 erroneous ideas of what to expect. In recording them we 

 must be in the highest degree unprejudiced and exact, and 

 neither add to nor detract from the truth ; for nature will 

 be sure to detect and expose any ' cooking of the results.' 

 We must also endeavour to notice every essential particu- 

 lar, and make an accurate and complete record of them at 

 the time, or at least before they have become at all dim in 

 the memory. In taking notes of experiments, it is of but 

 little use to record minor facts if the more important ones 

 are omitted; the investigator should therefore be pre- 

 viously qualified to appreciate in some degree the import- 

 ance of the more essential points. Generally also, as soon 

 as an experiment, or a few experiments, have been made, 

 conclusions should be drawn from the results, and sugges- 

 tions noted, because the particulars are then fresh in the 

 mind, and modifications can then be conveniently effected. 

 Effects which cannot then be understood may be left for 

 subsequent and more serious consideration. By a single 

 observation we usually add a single new fact to our know- 

 ledge ; but if that fact is an important one, like Oersted's 



