Light in a Diffraction Grating Spectrum. 399 



light incident at an angle of 4° 12' on the same side of the 

 normal as the spectrum. A bright line not much wider than 

 the distance between the D lines appears at wave-length 609, 

 and a dark band at 517 : the latter is sharp and black on one 

 side and shades off gradually on the other. On decreasing the 

 angle of incidence to 2° 37 / the bands approach, occupying 

 the positions shown in No. 2. 



Numbers 3 and 4 show two subsequent positions, and it will 

 be noticed that the rate o£ progress along the normal spectrum 

 is the same for each. In No. 4 we have the appearance 

 which I have likened to the line in the spectrum of a Nova. 



In No. 5 the incidence is normal and the lines have fused 

 and disappeared. This is not merely an approximation, for 

 I have found that if the grating be turned until the spectrum 

 has this appearance, the light reflected back through the 

 collimator passes through the slit. This furnishes us with 

 a new method for adjusting a grating for normal incidence. 

 On passing this position a narrow bright line appears which 

 broadens into a very sharply defined' rectangle, appearing, as 

 is shown in No. 6, at an incidence angle of 5' on the opposite 

 side of the normal from the spectrum. This rectangle 

 broadens as the angle of incidence increases, its edges be- 

 coming heavily shaded, as is shown in No. 7, where we 

 have essentially two dark bands retreating from each other 

 at equal rates as the angle of incidence increases. There 

 is nothing especially peculiar about the one which is journey- 

 ing towards the violet end of the spectrum, but the other 

 behaves in a most singular manner, which could only be fully 

 illustrated by a kinematograph-view of its changes. In No. 7 

 we find it very sharp and black on the right-hand edge, 

 shading off towards the red end of the spectrum. As it 

 moves along, a shadow appears on its right-hand side, the 

 two shadows being separated, however, by a narrow bright 

 region ; the right-hand shadow increases in depth, while the 

 left-hand one clears up, until the band becomes symmetrical, 

 a narrow bright line with a shadow on each side, as is shown 

 in No. 8. On increasing the angle of incidence still further, 

 the inverse of this operation takes place, until in No. 9 the 

 shadow has transferred its position to the right, and appears 

 with a sharp black edge as before only reversed in position. 

 This process of turning inside out of the shadow marks the 

 beginning of another curious event, for, as the reversed 

 shadow travels along towards the red with increasing angle 

 of incidence, an exceedingly black symmetrical band splits 

 off from it and travels down the spectrum in the opposite di- 

 rection, arriving at the position shown in No. 10, at an incidence 



