THE AID OF THE ACHROMATIC FRINGES. 



79 



there been no difference at any color, the horizontal fringes would have been 

 retained in that part of the spectrum. Such experiments may be made with 

 astonishing ease and accuracy, but this interpretation is involved. 



Now, as every lens is practically a plate for a small area near its optical 

 center, it is obvious that the same results must be obtained for a limit hori- 

 zontal band of the spectrum, if lenses are used as at g in figure 82. I tested 

 a concave spectacle-lens of about 2 diopters focal power, suspending it in the 

 solution from a pair of hard-rubber forceps, /, figure 83, slotted at the end to 

 receive the lens g. As one beam only passed through the lens, which was 

 here stronger in refractive index than the solution, the slit -images were not 

 in the same focal plane, but the corresponding parts were of the same height 

 in the field. On bringing the two slit-images into coincidence, the spectro- 



84 



I 



82 



telescope showed the fringe design in figure 84, the spectrum being" 1 clear 

 from violet v to red r, except at the longitudinal band ab. On raising and lower- 

 ing the lens in the solution, this band, ab, rose and fell correspondingly, and 

 when the lens was withdrawn the horizontal fringes at once appeared. As a 

 rule the band ab is not seen sharply in the principal focal plane and the ocular 

 must be drawn out or in, or an objective lens added. If the other beam (2 j, 

 fig. 82) is made to pass through the lens, the same phenomenon is observed, 

 except that the fringes are now upward toward the red. If the ocular is drawn 

 out in one case it will have to be drawn in for a sharp display of fringes in the 

 other. This seems to be due to the fact that the middle position (equidistant 

 from the half -silver mirror N) coincides with the mirror m', so that the lens 

 is not quite symmetric in either position. Sharp fringes occur when the area 

 of the washed slit-images (irrespective of sharpness of focus for slit-images) 

 coincide. 



59. Adjustable compensators. It is difficult to install a V-compensator 

 into either beam 8 5 or 2 i in figure 82, as the rays are usually too close to- 

 gether. A Billet glass wedge compensator suggests itself, but the introduction 

 of glass-path excess modifies the fringe number and would ultimately obscure 

 or wipe out the achromatic fringe. For the case of spectrum fringes, however, 

 it proved to be a very serviceable instrument. 



The air compensator (fig. 85) is also available. Here cc, c' are a pair of 

 narrow rectangular brass tubes, about i cm. wide, 2 cm. high, 10 or more 

 cm. long, closed at each end with the glass plates b and b'. Each glass plate 

 is common to both tubes, so that both beams 8 5 and 2 i may penetrate 



