rollers, balls

a b c d e

I say that if a weight of flat surface moves upon another like plane of dense surface, then to interpose between them the rollers is as easy for their motion as the balls. Nor do I see between these balls and rollers any difference except that the balls have universal motion, in every direction, and the rollers can travel only in one direction. But if the balls, or the rollers, touch one another in their motions, beyond comparison they will make the motion more difficult than if they were without contact. For their touchings rub together by contrary motions, as is shown in the 5th of the 6th, and by such rubbing they come to oppose the motions. But if the balls, or let us say the rollers, are somewhat apart, they will touch, between the weight and its resistance, only at a point, as is shown in the penultimate of the 2nd; whence it is as if they were supported upon the extremities of the line, so that it is an easy thing to give such movement.

I ask whether these four supports are to twist, and whether they will twist inward or outward. If you load the corners more, at top and at foot, on the inner side than on the outer, then the support will make its arch, that is convex, on the inner side. And loading from the outer side, the hollow of the arch will be made on the outer side. This that is said happens because the part that is loaded and pressed desires to draw together; whence it is necessary that the hump be made on the opposite side.

It is true that, through the inequality of the wood, it often does the opposite of its reason, for where the part of the wood is denser, there it resists. And so conversely, where it is less dense it resists less.