We do not forget that we are dealing with the shell of a hollow sphere, not with a ring, or section of a cylinder, and we can conceive that there would be, from the first, partial motions of translation in multitudes of directions, radial, angular, transverse, etc. etc., constantly changing, even being sometimes reversed, but also constantly combining with each other, and gradually leading on to decided, though partial, uniformity in one direction. As a matter of course this motion of translation would be controlled by its own constituent parts attracting each other to some extent, and thus a rotary motion would be established in the interior of the nebula in the region of greatest density. We can also conceive that when the motions of translation had become nearly uniform, the plane of that uniform motion might be in any direction through the whole mass of the nebula, and might be continually varying until final uniformity was attained, when the greater part of the mass was moving in combination, and the rotation was thereby firmly established in one direction, though still not embracing the whole. We have to take into account also that when the rotary movement had settled down into one plane, it would be most active at the distance of the region of greatest density of the nebula from its centre; in fact it would be instituted at that region and be, therefore, most active there; and then the most active part of the matter would be in the form of a rotating ring, still surrounded by an immense mass of nebulous matter, both inwards and outwards, to which it would gradually communicate its own motion, until the whole mass would rotate, in one direction, on an axis. But it is evident that in the whole rotating mass there would be different degrees of velocity of rotation at different places, decreasing from the supposed ring inwards towards the centre, and outwards to the surface at what would thus become the equatorial region; and also decreasing from the equatorial plane to the poles. Following up this idea, we have a more reasonable manner of accounting for the different velocities of rotation observed on the surface of the sun, between the equator and the poles, than we have seen suggested in any speculations on the cause that have come under our observation. Until rotation was fully instituted, the areolar law could have no power over the multitudinous movements going on in the nebula, but from that time it would begin to act, and condensation would increase it at the region where it began; and as all increase had to be propagated from there, inwards, outwards, and in all directions, the differences in velocity of rotation throughout the sun must endure as long as he continues to contract. In this we find an immense field for producing heat in the sun, from the eternal churning which must be going on in the interior.