The following is a brief account of a new apparatus for fine measurement in wavelengths of light, designed primarily as a comparator for the measurement in wavelengths of the difference between a standard of length, either a line or an end measure bar—the Imperial Standard Yard, for instance—and any duplicate or similar bar proposed to be employed as a derived standard. The instrument is also, however, the most perfect instrument yet devised for measurement in wavelengths in general, and performs its functions so admirably as to render it highly desirable that a description should now be published concerning it. It has been constructed to the designs and under the supervision of the author for the Standards Department of the Board of Trade, and this account of it is communicated to the Royal Society with the permission of the President of the Board of Trade. The principle underlying the instrument is that of the author’s interferometer, which has also proved so successful in its application, in the interference dilatometer, to the determination of the thermal expansion of small bodies by the Fizeau method, and in the elasmometer, to the measurement of the elastic bending of a small plate or bar under a given weight applied at the centre. The essence of the interferometer is that homogeneous light, of a definite wave-length, corresponding to a single spectrum line—isolated with the aid of a constant-deviation prism from the spectrum derived from a cadmium or hydrogen Geissler tube, or a mercury lamp—is directed by an autocollimation method, ensuring identity of path of the incident and reflected rays, normally upon two absolutely plane surfaces, arranged close to each other, and nearly, but not absolutely, parallel; the two reflected rays give rise, by their interference, to rectilinear dark interference bands on a brilliantly illuminated background in the colour corresponding to the selected wave-length. In the instrument now described, one of these two reflecting surfaces concerned in the production of the interference bands is carried by, and moves absolutely with, one of the two microscopes employed to focus the fiducial marks, or "defining lines", determinative of the length of the standard, the other surface being absolutely fixed. The movement of either of the surfaces with respect to the other causes the interference bands to move, and the extent of the movement of the surface is equal to half the wave-length of the light employed for every interference band that moves past a reference mark carried by the fixed surface. The movement of the microscope parallel to itself and to the length of the standard bar is thus measured by counting the number of bands and the initial and final fractions of a band which are observed to pass the reference spot during the movement, and multiplying that number by the half wave-length of the light radiation used in the production of the bands. It is only necessary, therefore, in order to compare the lengths of two bars, (1) to place the bar of known length, say, the Imperial Standard Yard, under the two microscopes so that the two defining lines are adjusted in each case between the pair of parallel spider-lines carried by each of the micrometer eye-pieces; (2) to replace the standard by the copy to be tested, so that the defining line near one end is similarly adjusted under the corresponding microscope, then, if the other defining mark is not also automatically adjusted under the second microscope which carries the interferometer glass surface, as it should be if it is an exact copy, (3) to traverse that microscope until it is so adjusted, and (4) to observe and count the number of interference bands which move past the reference spot during the process. The product of this number into the half wave-length of the light used to produce the bands thus obviously affords the difference between the two lengths included between the defining marks on the two bars.
The use of wave-length rulings as a defining lines on standards of length