26 Rlieinberg: Common Basis of the Theories ofMicroscopic Vision, XIX, 1. 



Now lastly we come to consider the result of a very large 

 inimber of sh:)ts, say some thousauds. We here get the bright bands 

 reduced in width to their lowest limits , tliey are eacli of them au 

 absolutely sharp Image of the slot used as an object. 



Though figure 29 already shows that any further increase of 

 slots would not alter the intensity curve much, as by the construction 

 there employed it would only meau addiug more lines of very small 

 height with greater numbers of maxima, there is auotlier and perhaps 

 still simpler way of regarding the matter. Suppose we have 1000 

 slots. Consider a point so exceedingly close to any maximum of 

 light, that waves reaching it from the first and second slot ditfer in 

 phase by only one-thousandth wave length. Then the waves reaching 

 that point from the !*'• and .501**- slot will dilfer by just ^ wave 

 length, and therefore neutralise one another. The same with the 

 2^"^- and 502 "'^•, and so forth. At very slightly dilfering slants 

 the 1^'and 500*^- the 2^^^- and 501^^- slot neutralise each other etc. 

 Thus everywhere except at the maxima themselves (at wbich points 

 waves arriving from any and every two slots differ by an exact 

 even number of wave lengths) , the effect of one slot will be neu- 

 tralised by the effect of some other slot, and therefore the bright 

 lines or Images of the slot are perfectly sharp. Already in figure 28 

 it will be observed that the contours of the central and first Image 

 on each side are fairly sharp Images of the irregulär slot which 

 served as the object. 



That the width of all the Images is nearly the same, is due 

 to the fact of the photo being taken with approximately monochromatic 

 light. ^ Just as with a Single slot, the maxima for dift'erent wave 

 lengths are formed at difterent points, and therefore spectra are 

 formed. Only that with a Single slot the spectra are impure by 

 reason of overlapping of broadened out bands formed by each wave 

 length (fig. 11), whereas now, the overlapping is very greatly di- 

 minished because the spectra consist of a series of sharp Images of 

 the slot corresponding to each wave length. We have in fact arrived 

 at the principle of the diffraction spectroscope. 



It may here be pointed out that the length or dispersion of 

 the spectra are directly proportionate to their distance from the 

 central Image for : 



') AU the photos (%. 15—17, 21—23, 27, 28) were taken through a 

 malachite green screen, passing a moderately narrow band of the spectrum 

 about the F line. 



