This is a paper which I wrote for Hamilton College's Mathematics 262: Geometries on the history of a geometric construction called the Butterfly Problem and explaining two proofs of its validity. I present one proof by elementary Euclidean geometry and one using analytic geometry. My paper is based off of an article entitled The Metamorphosis of the Butterfly Problem, written by Leon Bankoff, whose proof summaries I have expanded upon.
Figure 1:
The Butterfly Problem in its general case.
In a circle (O), P is the midpoint of chord AB.
Chords RS and TV pass through the point P.
RV cuts AP at a point M, and ST cuts PB at a point N.
Prove … that MP equals PN.
Figure 1 illustrates this construction.
In his article, The Metamorphosis of the Butterfly Problem,1 Leon Bankoff presented a series of proofs and a brief history of what has become known as the Butterfly Problem.
This simple arrangement has been around the mathematical world for about two centuries, according to Bankoff. Through the years, many mathematicians have tackled the Butterfly Problem, producing at least thirteen conceptually distinct proofs, which Bankoff summarized in his article.2
These approaches to the Butterfly Problem range from several elementary Euclidean geometric concepts, trigonometry, analytic geometry, geometric transversal theory, and other methods of treatment. All of this study resulted in a wealth of material proving the Butterfly Problem in almost any desired way.
Bankoff suggested that this great interest in the Butterfly Problem may come from its initial visual simplicity and its resemblance to the insect (from which the problem's name comes) which people find almost universally beautiful.3
Bankoff stated the Butterfly Problem explicitly (quoted from a proposal in a late 1950s edition of School Science and Mathematics) as:
In a circle (O), P is the midpoint of chord AB. Chords RS and TV pass through the point P. RV cuts AP at a point M, and ST cuts PB at a point N. Prove … that MP equals PN.
Figure 1 (above) illustrates this construction.
The earliest two proofs of the Butterfly Problem that Bankoff found appeared in an 1812 edition of The Gentleman's Diary of London. They take varied, but equally sound approaches to the problem, and the second even generalizes the positions of some of the points in the Butterfly Problem.
A brief study of the proofs that Bankoff presented confirms, as he suggested, that the most elegant proofs of this problem do not lie in the simplest mathematical approaches to solving it. The elementary Euclidean proofs are complex in setup and somewhat cumbersome to follow,4 whereas the more advanced proofs, though more complicated in material, are cleaner in presentation.
The first proof that I will present, Bankoff took from his own article published in the February 1955 edition of School Science and Mathematics.5
Figure 2: The Butterfly Problem as Bankoff originally visualized.
Note: This proof references Figure 2
Bankoff stated that this proof is also applicable to the case when the arc AR is smaller than the arc TR, so L would lie between T and B in Figure 2, by substituting the word equal for the word supplementary in each instance of his outline.
This method is not directly applicable to my expansion, as the words are used in more places than in the outline provided, but provides a good start to seeing where the proof 's adaptation would go.
Now that we have seen a solution by elementary Euclidean geometry, let us look at a proof using analytic geometry. This method, Bankoff notes, may be approached in two ways: one rather harsh and one more reasonable. The harsher method begins by setting the x-axis along AB and the y-axis along MO. Then one would find the equation of the circle and look for the x-intercepts of the lines FC and ED as defined by the intersections of CD and FE with the circle. Finally, one finds that the intercepts are equidistant and opposite from the origin. However, this method is very long and, as Bankoff said, it is a nasty method, recommended only for mathematicians with masochistic tendencies.
13
As an alternative, we will look at an analytic proof by Kesiraju Satyanarayana.14 This proof is concise and, in my opinion, elegant.
Figure 3:
The Butterfly Problem as Satyanarayana proved it.
Note: This proof references Figure 3 and Figure 4 (below).
Figure 4:
The intersecting conics of Satyanarayana's proof of the Butterfly Problem.
These proofs are but two of many ways that have been used to solve the Butterfly Problem through the years. The remaining eleven in Bankoff's article involve several other geometric approaches and a few from other disciplines. Some of them also deal with related theorems and, more relevantly, with generalizations of the Butterfly Problem to varying positions of M and other similar constructions. I recommend a study of the remaining proofs for an in-depth, though daunting, look at this problem.
If two triangles have two sides and the included angle of one congruent, respectively, to two sides and the included angle of the other, then the two triangles are congruent.(SAS) (Smart, page 410)
If a quadrilateral is inscribed in a circle, then the opposite angles are supplementary.(Smart, page 411)
If two parallel lines are cut by a transversal, then the alternate interior angles are congruent and the corresponding angles are congruent.(Smart, page 411)
A quadrilateral is inscribed in a circle if and only if the opposite angles are supplementary.(Smart, page 161)
An inscribed angle is measured by half its intercepted arc.(Smart, page 411)
If two triangles have two angles and the included side of one congruent, respectively, to two angles and the included side of the other, then the two triangles are congruent(ASA) (Smart, page 410)
I would like to thank Sarah Britton for the comments in her review of this paper and Jess Kapps for her proofing of later drafts.