About the mass there has been, and still is, great uncertainty. This is because it can only be found from the perturbing effect it has on Venus, the Earth, or Encke’s comet. Modern determinations, however, are smaller than the older ones; thus Backlund in 1894 got from the effect on Encke’s comet only one-half the mass that Encke had, fifty-three years before. Probably the most reliable information comes from Venus, which Tisserand found to give for Mercury ¹/₇₁₀₀₀₀₀ of the mass of the Sun, or ¹/₂₁ of the mass of the Earth. If we take ¹/₇₀₀₀₀₀₀ as the nearest round number, we find the planet’s density to be 0.66 that of the Earth. [Illustration: MAP OF MERCURY LOWELL OBSERVATORY 1896-97] The same observations that disclosed at Flagstaff the planet’s size revealed a set of markings on his face so definite as to make the rotation period unmistakable. It takes place, as Schiaparelli found, in eighty-eight days, or the time of the planet’s revolution round the Sun. The markings disclosed the fact, as Schiaparelli had also discovered, in a most interesting manner, for the ellipticity of the planet’s orbit stood reflected in the swing of the markings across the face of the disk, a definiteness in the proof of a really surprising kind. What this means we shall see in a subsequent chapter when we take up the mechanical problem of the tides. Another result that issued from the positions of the markings was the determination of the planet’s pole. Except for the libration above noticed, the markings kept an invariable longitudinal position upon the illuminated disk, showing that the planet turned always the same face to the Sun; but latitudinally a difference was noticeable between their place in October-November, 1896, and in February-March, 1897, the latter being 4° farther north. Now this is just what the orbital position should have caused, if the pole stood vertically to it. Thus a difference of 4° from perpendicularity should have been discernible, had it existed,—a very small amount in such a determination. We may, therefore, conclude that the axis stands plumb to the orbit, and this is what theory demands.