248 ON THE THEORY OF THE MICUOSCOPE. 



of the total image do not blend together correctly, but will be 

 thrust out laterally or backwards. Marks of structure which 

 belong in the object to one and the same place and level {e.g.^ 

 various systems of lines) appear separated from each other, and 

 also from the points in the object to which they belong. In 

 consequence of the onesidedness with which, in modem times, 

 the improvement of the microscope has been directed towards 

 the increase of angular aperture, the conditions under which 

 abnormal appearances, and especially deceptive alterations of 

 level are produced, occur abundantly in the new high power 

 objectives, as repeated experience has shewn me, and 

 I assuredly do not err in expressing my conviction that 

 the consequences of this state of things affect to an 

 unexpected extent the numerous questions in dispute 

 amongst microscopists, concerning the interpretation of minute 

 structures. 



Since every one must admit that the first and most imperative 

 claim which can be made, in the interest of scientific microscopy, 

 upon the performing power of the instrument is this — that parts 

 which belong together in the object shall also appear as belonging 

 together in the microscopic image, it follows that uniform cor- 

 rection of spherical aberration throughout the whole area of 

 aperture must be the absolute criterion and rule of guidance in 

 the construction of a microscope. Now, it has been shewn 

 (paragraph vii.)that with a dry objective an adequate compensation 

 of spherical aberrations is, as a matter of fact, impossible when the 

 angular aperture exceeds 110°. Hence it must be concluded that 

 a dry objective will be less suited for ordinary scientific use in 

 proportion as it renders visible such finer systems of lines as 

 exceed the limits of resolving power answering to that angle 

 (namely, 0.3o/x for oblique light). The greatest possible increase 

 of resolving power can be obtained in a rational way only by 

 means of immersion objectives, as these alone admit of the 

 largest possible {i.e. technically practicable angular) aperture, 



