Biola University Biola Home Page
My Math Papers
Woopy Home Page

A New Proof that the Spieker Point is the Radical Center of the Excircles of any Triangle.

Peter Y. Woo, 11/11/98


    Let AA', BB’, CC' be the medians of DABC and I' the in-center of DA'B'C'. Let a, b, c denote the lengths of BC, CA, AB, and s = (a + b + c)/2. Let the excircles KA, KB, KC have centers IA, IB, IC and radii rA, rB, rC. Let r, R be the inradius and circumradius of DABC. Without loss of generality, assume c > a > b.

    Theorem 1. The radical center of the 3 excircles of DABC is the Spieker point of DABC, namely, the incenter I' of DA'B'C'.
    Proof. Let the lines C'I' and AC intersect at S. We shall prove C' and S are midpoints of common tangents of KA and KB. First let ZA and ZB be the feet of perpendiculars from IA and IB to line AB. Then AZB = BZA = s - c, hence C' is the midpoint of ZAZB. Next, let C'I'S cut A'B' at V, and let YA, YB be feet of perpendiculars from IA and IB to line AC. Then SB': B'A = B'V : AC'-B'V' = B'V : B'A'-B'V = B'V: A'V = B'C': A'C' = a : b. Hence SB' = b/2 × a/b = a / 2. Then SC = (a-b) / 2. From CYB = s - a and CYA = s - b we can deduce S is the midpoint of YAYB.
    Since C' and S are midpoints of common tangents of KA and KB, C'I' is the radical axis of KA and KB. Similarly B'I' is the radical axis of KA and KC. Hence I' is the radical center of the 3 excircles. QED


    Theorem 2. Let B'A' be extended to P and C'A' be extended to Q so that A'P = A'Q = a / 2. Then P, Q lie on the circle centered at I' orthogonal to the 3 ex-circles of DABC . Also ZAPQYA is a straight line.
    Proof. Let DABC have angles 2 a, 2 b, 2 g at A, B, C. Let N be the foot of perpendicular from I' to A'B'. Then I'P2 = PN2 + I'N2 = (a/2 + (s-a)/2)2 + (r/2 )2 =(s2 + r2)/4. I'P = I'Q is also trivial. So P,Q lie on the circle K' centered at I' with radius Ö(s2 + r2) / 2.
    Next, we need to prove I'IA2 -I'Q2 = rA2 to show that K' is orthogonal to KA. Let C'S cut IAC at T. Notice C'T is perpendicular to IAC [either because C'I'TS is the radical axis and so is perpendicular to IAIB, or from computing angles]. Hence I'IA2 -I'Q2 = IAT2 -QT2 = (IAQ+QT)2 -QT2 = IAQ2 + 2 IAQ . QT .
    Now C'Q = b/2 + a/2. \QT = (½)(a + b) sin g . Again, since A'PQ is isosceles, PQ = 2 . (a/2) . sin a = a sin a. By angles chasing, we can see DPQIA has angles g‘= p/2 -g, b‘=p/2 -b, a‘=p/2 -a respectively \IAQ = PQ sin g‘ / sin a‘ = a tan a cos g. It follows that
    IAQ2 + 2 IAQ . QT = a tan a cos g (a tan a cos g + 2 (½)(a + b) sin g ) . A most refreshing trigonometry ordeal will simplify this to s2 tan2a = rA2.
    [Here are my proud details:
    a tan a cos g (a tan a cos g + (a + b) sin g )
= a tan a sec a cos g (a (sin a cos g + cos a sin g ) + b cos a sin g )
= a tan a sec a cos g (a cos b + b cos a sin g )
= a tan a sec a cos g ((4 R sin a cos a) cos b + (4 R sin b cos b) cos a sin g )
= a tan a cos b cos g 4 R ( sin a + sin b sin g )
= a tan a cos b cos g 4 R cos b cos g
= tan2a (4 R cos a cos b cos g)2
= s2 tan2a .]

    To see that ZAPQYA is a straight line, observe that C'ZA = c/2 + (s - c) = (a + b)/2= C'Q, so that
ÐC'QZA = (p - 2a)/2 = ÐC'QP. Hence ZAPQ is a straight line. Similarly PQYA is a straight line. QED

    Acknowledgment. Thanks to Paul Yiu of Florida Atlantic Univ. for directing me to these theorems.

Home Site Map Search Biola Feedback Home

Direct comments or questions to: Woobiola, woobiola@yahoo.com