264 
due the credit of first showing the true nature of the con- 
ditions on which the resolution of minute structure depended. 
Aided by Mr. Stephenson and Mr. Frank Crisp, chief editor 
of the Journal of the Royal Microscopical Society, he 
devised a series of simple, and therefore ingenious, experi- 
ments by which it was demonstrated that the resolution of 
minute structure is essentially dependent on two factors, 
namely, (1) the capability of the structure to diffract the 
light by which it is illuminated ; (2) on the capacity of the 
objective to receive and bring to a focus these diffracted 
rays. Before proceeding to show the very conclusive ex- 
periments by which Professor Abbe demonstrated the truth 
of the above assertion, I may, perhaps, be allowed to say a 
few words on the nature of the diffraction of light, taking, 
as an example, the diffraction due to a series of very narrow 
lines or grooves, such as is seen on the scales of butterfliesj 
Lepisma saccharina, &c., and, more especially, in the arti- 
ficially ruled glass and metal plates, known as diffraction 
gratings. 
If a sunbeam, passing through a very small opening, be 
allowed to fall on such a structure, and after passing through 
the transparent interspaces, is made to pass through a lens, 
and so to form a clear image on a screen, this image is seen to 
consist of a central colourless representation of the opening, 
with a series of spectra arranged symmetrically on either 
side, with their violet ends towards the central colourless 
image. If the grooves of the grating are very near together, 
say the slo mm,, the spectral images are widely separated 
not only from each other, but also from the central colour- 
less images, and the dispersion of each spectrum is propor- 
tionately great. When, therefore, the distance from the 
centre of one clear interspace to that of another is less than 
that of a wave length of light, the deviation of even the 
nearest spectra to the central image becomes so great as to 
become an imaginary quantity, and no lateral spectra can 
