John William Strutt, third baron Rayleigh, O.M., F.R.S., sometime president of the Royal society and chancellor of the University of Cambridge / by his son Robert John Strutt, fourth baron Rayleigh.

Robert Strutt, 4th Baron Rayleigh

Date:: 1924

Credit: John William Strutt, third baron Rayleigh, O.M., F.R.S., sometime president of the Royal society and chancellor of the University of Cambridge / by his son Robert John Strutt, fourth baron Rayleigh. Source: Wellcome Collection.

118/434 (page 96)

$doubled. The angular separation between the diffracted images of the two components would be halved 3 but as against this, the images themselves would be of only half the breadth, and therefore the grating would still remain just adequate to the work.1 The argument as given applies only to the first diffracted image or spectrum* It will, however, easily be seen that the resolving power in the second spectrum will be twice, and in the third spectrum three times, as great. These latter images are as narrow as the first ones, and the angular separation between the component soda lines is 30 seconds of arc for the second spectrum, and 45 seconds of arc for the third spectrum. Here is the entry in Rayleigh’s notebook in which these conclusions were first put to the test, Sept. 8th, 1873. “ Made a rough estimate by observation of the number of lines required to resolve the soda lines in the various spectra. The 3,000 original was used, but did not seem to define better than a copy, though it required a different inclination of the object lens. The widths of slits required were put at 25, 42, 79 200ths of an inch, which should be in the ratio 2:3:6. If we take 8 200ths as required in the 1st spectrum, we find for the number of lines x 3,000 = 1,200. “ I had supposed from theory that 1,500 or 1,000 might be required : but * resolution ’ is of course a vague thing.” It was not until five years later that he completed the subject by investigating the more ordinary form of spectro¬ scope in which a prism is used to form the spectrum instead of a grating. He wrote:— “ At the time the above paragraph [on resolving power of grat¬ ings] was written, I was under the impression that the dispersion in a prismatic instrument depended on so many variable elements that no simple theory of its resolving power was to be expected. Last autumn, when engaged upon some experiments with prisms, 1 The observing telescope would, of course, have to be made larger to cover the larger grating. Hence a close ruling of the grating is generally advantageous in practice.$