XXXIII. ‘Bravo, Theodorus!’ said Ammonius; ‘he has worked out his task with spirit. Yet I shall be surprised if he is not found to be using assumptions which are mutually destructive. He wants to have it that all five solids do not attain their structure together, the finest and easiest of [Sidenote: F] composition always breaking first into being. Then, as though following upon this, and not conflicting with it, he lays it down that it is not all matter which admits the finest and simplest element first, but that sometimes the heavy and complex objects are the first to emerge out of matter. But to pass over this, whereas it is assumed that there are five primal bodies, and therefore an equal number of worlds, he makes out probability for four only; he has discarded the Cube as if playing at [Sidenote: 428] counters, since Nature does not adapt it to pass into the others or them into itself, because the triangles are not of the same kind. In the other cases the basis of formation is the semi-(equilateral) triangle; to the Cube the right-angled isosceles is peculiar, which is incapable of converging towards the others or joining with them to form one solid angle. If then, there are five bodies and five worlds, and in each the primary belongs to some one, then where the Cube has been the first to come into being, the rest will be nowhere, for it is unable to change into any of them. I pass by the fact that they make the element of the Dodecahedron also a different thing from that scalene [Sidenote: B] out of which Plato constructs the Pyramid, the Octahedron, and the Eicosahedron. And so’, added Ammonius, with a laugh, ‘you must either resolve these difficulties, or give us something of your own about the common problem.’