Preface

Lorenzo Mascheroni (The Geometry of the Compass, Pavia 1797, pp. 3–8)

The first thought that invited me to explore the new paths of this Geometry of the Compass was this: while so many new truths are discovered by progressing forward in mathematics, might not some unknown territory be found by stepping backward? Hitherto, the simplest solutions in geometry have been judged to be those employing nothing other than the compass and the straightedge; or, what is the same, the straight line, which is the simplest among lines, and the circle, which is the simplest among curves. To these two instruments of geometric problems, which formerly bounded and constituted elementary geometry, conic sections were subsequently added; then higher curves and transcendents of various kinds. These profound geometric inquiries continued to be enriched by the new resources of algebra, both finite and infinitesimal, to such a degree that discoveries which initially appeared marvelous to the ancients, and worthy of the sacrifices of Thales and Pythagoras, have now become the common possession of children in our schools.

Now I asked myself: could you not step backward from the elements, as from a boundary line, and seek something that has remained behind, as if overlooked? Is it truly certain that the elementary problems of Euclid are of the simplest construction? Or might the mathematical element itself be decomposed into its fundamental components, the straightedge and the compass—in the manner of a chemist who has separated water into two airs, and even an air once deemed simple into two other substances? At this point I perceived that while the straightedge alone can do nothing except draw a straight line between two given points, the compass alone might perhaps be employed not merely to describe a single circle or circular arc, but by describing several circles with several centers and diverse openings, to locate by their mutual intersections multiple points that would be useful, and precisely those whose positions were sought in a given problem.

Up to this point I knew that this was a branch hitherto entirely uncultivated by mathematicians, and that solutions of this kind obtained by the compass alone would be, in their construction, more elementary than any others. Yet two thoughts briefly held me back from undertaking anything in this direction. The first was: what benefit will accrue if you succeed in finding with the compass alone those points which others have already found with it and the straightedge? The second was the apprehension—which at first seemed thoroughly reasonable—that far from gaining an advantage from my attempts, even if crowned with success, I should suffer a loss: constructions with the compass alone to determine the points of elementary geometry would be multiplied manifold beyond those already known in which the straightedge is employed; theory would thus lack elegance, and practice would lack precision. Hence I was on the verge of abandoning the enterprise.

While I remained thus irresolute, it happened that I re-read the method by which Graham and Bird divided their large astronomical quadrants in England (Encyclop. Method. Articl. quart de cercle mural). Graham’s quadrant, constructed by him for Greenwich, is said not only to have served as the model for most of those made subsequently, but is still regarded by astronomers for its precision as one of the best ever employed in astronomy prior to the era of Ramsden’s quadrants. Now I observed that the graduation of that celebrated instrument, the straightedge having been completely discarded, was executed with the compass alone. The description of the method employed in that lengthy and ingenious operation is remarkable. I shall not enter here into the reasons why the straightedge was excluded; all who possess practical skill in such work will easily divine them. To indicate in general the advantages which the compass possesses over the straightedge when precise delineation of lines is required that must not fear microscopic examination, it suffices to observe that when dealing with a straightedge of considerable length, it is virtually impossible that it should be so perfectly straight as to guarantee throughout its entire span the true location of points placed along it. And even granting it to be absolutely straight, practical artisans know well that dragging a scribing point along its edge introduces an uncertainty of parallelism in the motion of the scriber’s axis, or an imperfect adherence to the guide, which often frustrates the highest perfection. To these two hazards the compass is not subject: provided its span remains rigid and its points are exceedingly fine, once one point is fixed immovably—which is not difficult—the other, as it sweeps, traces of itself an arc so precise and true that nothing can surpass it.