A new method of obtaining refractive index was described in 1902. Instead of depending on the accuracy of arbitrary measurements of refracting angle or the accuracy of construction of a prism, this method was based on the mean value of the three angles of a triangle. The description of this method was accompanied by a Table of the principal refractive indices of quartz, calcite, and fluorite, obtained in this manner, and was followed later by similar Tables of the refractive indices of vitreous silica, water, and optical glass, and in February, 1915, these were in turn followed by a Table of the refraction temperature coefficients of optical glass. Since the error due to the method itself amounted to less than unity in the seventh figure, there was promise of great accuracy. But, nevertheless, with every care, and in spite of the correction for temperature, errors even exceeding unity in the fifth decimal place were frequently manifest, requiring repeated and tedious observation to eliminate. Hence suspicion fell on the barometer. But it was pointed out by the late Prof. Sylvanus Thompson, F. R. S., and afterwards by Sir Arthur Schuster, who indicated the lines to be worked on, that the modulus of rigidity of glass precluded its being sensibly affected by atmospheric pressure, and that, therefore, any such effect must be due to the refraction of air alone.