THE METHOD OF SCIENCE 51 



at once: "I was thinking it was very bad." W'e were, of 

 course, ol^serving different things. He was interested in the 

 general appearance of the radiograph and would have been 

 equally pleased ^vith any photograph of the same subject. I 

 was critically observing the rendering of detail in the shadows, 

 in ^vhich respect that particular photographic material ^vas 

 unsatisfactory. 



Observations must be controlled by knowledge of the 

 errors which the sense organ itself may introduce in the ob- 

 servation. The W'hole class of optical illusions, for instance, 

 may produce false conclusions. The unaided ear, and espe- 

 cially the untrained ear, cannot be trusted to give reliable 

 information as to sounds. There is also the question of 

 personal error. The observer must recognize what H. G. 

 Wells calls "the limitations of the instrument," not only as 

 regards the sense organ but also, as Wells uses it, in regard 

 to the mind itself. 



In scientific research, observation is not always direct; 

 much use is made of instruments and apparatus. Instead of 

 the eye, the photographic film or the photoelectric cell may 

 be used. Sound vibrations may be measured electrically. 

 Instruments have many advantages over the unaided senses. 

 The microscope makes very small things visible. The tele- 

 scope collects light from a large lens surface and then enables 

 magnification to be applied. Moreover, such instrumental 

 methods of observation enable us to overcome the limitations 

 imposed by the recording system of the brain. It does not 

 matter how unexpected or ho^v rapid and transient a phe- 

 nomenon is, if we have a photographic record of it. A 

 motion picture of the disturbance in the classroom ^vould 

 have enabled all observers to agree on the facts after they had 

 seen it several times. The sudden flash of the lines in the 

 spectrum at the second contact point of an eclipse can be 

 recorded photographically and studied at leisure. 



Observations made with instruments are essentially judg- 

 ments of coincidence. The observer measures a length by 

 seeing the point at w^hich the object to be measured comes 



