1886.] 



MICKOSCOPICAL JOURNAL. 



191 



imaged back of the objective, in 

 its upper focal plane, and may be 

 seen by removing the ocular and 

 looking dow^n the tube of the ini- 

 croscope. 



By the combination of the spectral 

 images, which are images of the source 

 of light, or of the diaphragm, opening, 

 in the conjugate focal plane of the ob- 

 ject, the image of the refracting ele- 

 ments is produced, by interference. 



The closer the lines the greater w^ill 

 be the number of diffraction spectra. 

 When we observe a lighted candle 

 through a diffraction plate the closer 

 the lines the more images will be 

 seen. It wall be obvious, therefore, 

 that since the portrayal of the struc- 

 ture depends upon the gathering in 

 of the diffraction spectra by the ob- 

 ject-glass, it is important that all the 

 diffraction spectra should be so taken 

 up, for each series of spectra will pro- 

 duce a definite number of lines in the 

 image, and no more, independently of 

 the structure of the object. The num- 

 ber of spectra that an objective will 

 collect, the successive spectra being 

 formed further and further from the 

 optic axis, will depend entirely upon 

 the angular aperture of the lens. We 

 are thus able to understand the value 

 of angular aperture, and we see at 

 once why it is that resolving power 

 increases with angular aperture. 



The spectral images portray only 

 the minute structure of an object. In 

 addition to this we have the images of 

 grosser parts formed by the ordinarv 

 dioptric action of the lenses. The 

 skill of the maker is severely tested 

 to bring the dioptric and diffraction 

 images into the same plane. In the 

 resolution of a diatom frustule, such 

 as you will see this evening, we have 

 the dioptric image of the outline and 

 the central longitudinal line and the 

 diffraction images of the cross mark- 

 ings. In the case of Nobert's bands 

 of lines ruled on glass there is no di- 

 optric image. 



It results from the facts stated above 

 that the images of minute structures 

 seen in the microscope are interfer- 



ence images, and are, to a certain ex- 

 tent, independent of the details of the 

 structures under examination. In 

 other words, whatever elements will 

 give identical diffraction spectra will 

 be portrayed as identical structures. 

 Moreover, in the case of bands of 

 lines, by excluding certain spectral im- 

 ages and admitting others, the num- 

 ber of lines in the image, supposing 

 the object to be a band of ruled lines, 

 may be doubled. Various other modi- 

 fications may be made in the image 

 which time does not permit me to 

 mention. 



Having thus reviewed the present 

 theory of microscopic vision in a very 

 superficial manner, it remains to con- 

 sider the improvement in the con- 

 struction of microscope objectives 

 which the theory has led to. The 

 greatest improvement of late years 

 has been the adoption of a system 

 known as homogeneous immersion, 

 in which the front lens of the objec- 

 tive is brought into optical contact 

 with the object or the cover-glass by 

 means of an immersion ffuid having 

 an index of refraction the same as 

 glass. It is assumed that rays from 

 the object pass without refraction from 

 the object to the objective. With 

 such lenses the angidar cone of rays 

 entering the front lens is much smaller 

 than that entering a lens without an 

 immersion medium, nevertheless, a 

 greater number of diffraction spectra 

 will be taken in by such a lens. 



Owing to the effect of the immer- 

 sion media, it is evident that while 

 increase of angular aperture in any 

 medium gives greater power of reso- 

 lution, the same result may be at- 

 tained by reducing the angular aper- 

 ture and the use of an immersion 

 medium of higher refractive power. 



Therefore, the term angular aper- 

 ture is not sufficiently definite for 

 practical purposes, and Prof. Abbe 

 has introduced the term numerical 

 aperture, which is the product of the 

 index of refraction of the medium 

 multiplied by the sine of half the 

 angular aperture in that medium, n 



