This leads us to another significant fact concerning the peculiar condition of Jupiter's surface. Not only does the south belt move perceptibly faster than the red spot, but, generally speaking, the various markings on the surface of the planet move at different rates according as they are nearer to or farther from the equator. Between the equator and latitude 30 deg. or 40 deg. there is a difference of six minutes in the rotation period--i.e., the equatorial parts turn round the axis so much faster than the parts north and south of them, that in one rotation they gain six minutes of time. In other words, the clouds over Jupiter's equator flow past those in the middle latitudes with a relative velocity of 270 miles per hour. But there are no sharp lines of separation between the different velocities; on the contrary, the swiftness of rotation gradually diminishes from the equator toward the poles, as it manifestly could not do if the visible surface of Jupiter were solid. In this respect Jupiter resembles the sun, whose surface also has different rates of rotation diminishing from the equator. Measured by the motion of spots on or near the equator, Jupiter's rotation period is about nine hours fifty minutes; measured by the motion of spots in the middle latitudes, it is about nine hours fifty-six minutes. The red spot completes a rotation in a little less than nine hours and fifty-six minutes, but its period can not be positively given for the singular reason that it is variable. The variation amounts to only a few seconds in the course of several years, but it is nevertheless certain. The phenomenon of variable motion is not, however, peculiar to the red spot. Mr. W.F. Denning, who has studied Jupiter for a quarter of a century, says: "It is well known that in different latitudes of Jupiter there are currents, forming the belts and zones, moving at various rates of speed. In many instances the velocity changes from year to year. And it is a singular circumstance that in the same current a uniform motion is not maintained in all parts of the circumference. Certain spots move faster than others, so that if we would obtain a fair value for the rotation period of any current it is not sufficient to derive it from one marking alone; we must follow a number of objects distributed in different longitudes along the current and deduce a mean from the whole."[10] [Footnote 10: The Observatory, No. 286, December, 1899.]