I mean by this that, by writing it backwards, from its found ends {practical machinable teeth) to its tentative beginnings (dimly perceived geometrical notions), one might conceivably write a manual, not on how to understand these kinds of gears, but on how to make them.