MICROSCOPIC VISION. 157 



which the images of minute details depend. From this 

 point of view, it must be evident that the attempt to 

 determine the structure of the finer kinds of diatom valves 

 by morphological interpretation of their microscopic ap- 

 pearances is based on inadmissible premises. Whether, for 

 example, Pleurosigma angiilatum possesses two or three seta 

 of striae ; whether striation exist at all ; whether the visible 

 delineation is caused by isolated prominences or depressions, 

 etc., no microscope, however perfect, no amplification, how- 

 ever magnified, can inform us." 



This passage means that when you are dealing with 

 objects not less than -^^q inch in size, the microscope image 

 is to be trusted ; but with regard to structures of a smaller 

 size than this, because they are imaged diffractionally, we 

 have no warrant in asserting them to be morphologically 

 accurate. Now, the experiments given above show that the 

 diffraction limit is at least fifty times greater than that 

 assumed ; the figures, therefore, in the preceding paragraph 

 must be altered to -^^ inch, from which it follows that the 

 only microscope images to be accepted as truthful are those 

 of objects larger than the -^^ inch : a conclusion which we 

 know to be absurd. 



There was another conclusion derived from the diffraction 

 theory which was fraught with far greater danger because 

 it introduced erroneous methods of working the microscope. 

 It was said, because the image of fine structure depends 

 upon spectra, therefore make spectra. 



By this not altogether clear phrase, it was meant that the 

 spectra which can be readily seen when the eyepiece is re- 

 moved should be made as bright as possible by reducing the 

 aperture of the illuminating cone. E educing the aperture 

 of the illuminating cone is analogous in its effect to closing 

 the slit of a spectroscope, for if the slit is too wide there is 



