For a convex quadrilateral with sides a , b , c , d {\displaystyle a,b,c,d} , diagonals e {\displaystyle e} and f {\displaystyle f} , and g {\displaystyle g} being the line segment connecting the midpoints of the two diagonals, the following equations holds:
If the quadrilateral is a parallelogram, then the midpoints of the diagonals coincide so that the connecting line segment g {\displaystyle g} has length 0. In addition the parallel sides are of equal length, hence Euler's theorem reduces to
which is the parallelogram law.
If the quadrilateral is rectangle, then equation simplifies further since now the two diagonals are of equal length as well:
Dividing by 2 yields the Euler–Pythagoras theorem:
In other words, in the case of a rectangle the relation of the quadrilateral's sides and its diagonals is described by the Pythagorean theorem.1
Euler originally derived the theorem above as corollary from slightly different theorem that requires the introduction of an additional point, but provides more structural insight.
For a given convex quadrilateral A B C D {\displaystyle ABCD} Euler introduced an additional point E {\displaystyle E} such that A B E D {\displaystyle ABED} forms a parallelogram and then the following equality holds:
The distance | C E | {\displaystyle |CE|} between the additional point E {\displaystyle E} and the point C {\displaystyle C} of the quadrilateral not being part of the parallelogram can be thought of measuring how much the quadrilateral deviates from a parallelogram and | C E | 2 {\displaystyle |CE|^{2}} is correction term that needs to be added to the original equation of the parallelogram law.2
M {\displaystyle M} being the midpoint of A C {\displaystyle AC} yields | A C | | A M | = 2 {\displaystyle {\tfrac {|AC|}{|AM|}}=2} . Since N {\displaystyle N} is the midpoint of B D {\displaystyle BD} it is also the midpoint of A E {\displaystyle AE} , as A E {\displaystyle AE} and B D {\displaystyle BD} are both diagonals of the parallelogram A B E D {\displaystyle ABED} . This yields | A E | | A N | = 2 {\displaystyle {\tfrac {|AE|}{|AN|}}=2} and hence | A C | | A M | = | A E | | A N | {\displaystyle {\tfrac {|AC|}{|AM|}}={\tfrac {|AE|}{|AN|}}} . Therefore, it follows from the intercept theorem (and its converse) that C E {\displaystyle CE} and N M {\displaystyle NM} are parallel and | C E | 2 = ( 2 | N M | ) 2 = 4 | N M | 2 {\displaystyle |CE|^{2}=(2|NM|)^{2}=4|NM|^{2}} , which yields Euler's theorem.3
Euler's theorem can be extended to a larger set of quadrilaterals, that includes crossed and nonplaner ones. It holds for so called generalized quadrilaterals, which simply consist of four arbitrary points in R n {\displaystyle \mathbb {R} ^{n}} connected by edges so that they form a cycle graph.4
Lokenath Debnath: The Legacy of Leonhard Euler: A Tricentennial Tribute. World Scientific, 2010, ISBN 9781848165267, pp. 105–107 /wiki/ISBN_(identifier) ↩
Deanna Haunsperger, Stephen Kennedy: The Edge of the Universe: Celebrating Ten Years of Math Horizons. MAA, 2006, ISBN 9780883855553, pp. 137–139 /wiki/ISBN_(identifier) ↩
Geoffrey A. Kandall: Euler's Theorem for Generalized Quadrilaterals. The College Mathematics Journal, Vol. 33, No. 5 (Nov., 2002), pp. 403–404 (JSTOR) https://www.jstor.org/stable/1559015 ↩