Gaussian Guesswork ....... .... or Why 1.19814023473559220744 . . . is Such a Beautiful Number Article Author: Adrian Rice Math Horizons, November 2009, pp. 12 — 15. DOI 10.4169/194762109X476491. “Before theorems are proved, conjectures must be made, . . . ” “Before theorems are proved, conjectures must be made, . . . ” . . . and for that to happen, all kinds of experimentation, observation, invention and, indeed, imagination must come into play.” Carl Friedrich Gauss (1777-1855) • One of the most creative mathematicians of all time (alongside Archimedes and Newton). Carl Friedrich Gauss (1777-1855) • One of the most creative mathematicians of all time (alongside Archimedes and Newton). • Kept a “mathematical diary” for nearly 20 years (just before his 19th birthday in 1796 until July 1814) Gauss’ May 30, 1799 Discovery: Amazing relationship between three particular numbers: • a sophisticated form of average • a particular value of an elliptic integral • the ratio of the circumference of a circle to its diameter (π) Some of what Gauss knew about π • π ∼= 3.14159265358979323846etc. Some of what Gauss knew about π • π ∼= 3.14159265358979323846etc. • π = 2 ∫ 1 0 dt √ 1 − t2 (arclength of semi-circle) Some of what Gauss knew about π • π ∼= 3.14159265358979323846etc. • π = 2 ∫ 1 0 dt √ 1 − t2 (arclength of semi-circle) • If u = ∫ x 0 dt √ 1 − t2 , then x = sin u . . . Some of what Gauss knew about π • π ∼= 3.14159265358979323846etc. • π = 2 ∫ 1 0 dt √ 1 − t2 (arclength of semi-circle) • If u = ∫ x 0 dt √ 1 − t2 , then x = sin u . . . . . . and sin(u + 2π) = sin u. Hmmmmmm . . . . . . what happens with similar integrals????? If u = ∫ x 0 dt √ 1 − tn , then x = . . . January 1797: Begins to investigate case of n = 4: ∫ x 0 dt √ 1 − t4 January 1797: Begins to investigate case of n = 4: ∫ x 0 dt √ 1 − t4 Cool! Arclength of lemniscate curve: (x2 + y2)2 = x2 − y2 January 1797: Begins to investigate case of n = 4: ∫ x 0 dt √ 1 − t4 Cool! Arclength of lemniscate curve: (x2 + y2)2 = x2 − y2 Follow the sin u pattern to define a leminiscate sine!!! • If u = ∫ x 0 dt √ 1 − t2 , then x = sin u. • If u = ∫ x 0 dt √ 1 − t4 , then x = slu. • If u = ∫ x 0 dt √ 1 − t2 , then x = sin u. sin(u + 2π) = sin u , where π = ∫ 1 0 dt √ 1 − t2 . • If u = ∫ x 0 dt √ 1 − t4 , then x = sl u. • If u = ∫ x 0 dt √ 1 − t2 , then x = sin u. sin(u + 2π) = sin u , where π = ∫ 1 0 dt √ 1 − t2 . • If u = ∫ x 0 dt √ 1 − t4 , then x = sl u. sl (u + 2ω) = sl u , where ω = 2 ∫ 1 0 dt √ 1 − t4 . • If u = ∫ x 0 dt √ 1 − t2 , then x = sin u. sin(u + 2π) = sin u , where π = ∫ 1 0 dt √ 1 − t2 . • If u = ∫ x 0 dt √ 1 − t4 , then x = sl u. sl (u + 2ω) = sl u , where ω = 2 ∫ 1 0 dt √ 1 − t4 . ω = 2.62205755429211981046 etc. Gauss’ May 30, 1799 Discovery: Amazing relationship between three particular numbers: • the ratio of the circumference of a circle to its diameter: π Gauss’ May 30, 1799 Discovery: Amazing relationship between three particular numbers: • the ratio of the circumference of a circle to its diameter: π • a particular value of an elliptic integral: ω = 2 ∫ 1 0 dt √ 1 − t4 Gauss’ May 30, 1799 Discovery: Amazing relationship between three particular numbers: • the ratio of the circumference of a circle to its diameter: π • a particular value of an elliptic integral: ω = 2 ∫ 1 0 dt √ 1 − t4 • a sophisticated form of average The Arithmetic-Geometric Mean of a, b Define two sequences: n = 0 : a0 = a b0 = b n = 1 : a1 = 1 2 (a0 + b0) b1 = √ a0b0 n ≥ 1 : an = 12 (an−1 + bn−1) bn = √ an−1bn−1 Facts about (an), (bn) for a, b ≥ 0 • Both sequences converge. Facts about (an), (bn) for a, b ≥ 0 • Both sequences converge. • lim n→∞ an = lim n→∞ bn Facts about (an), (bn) for a, b ≥ 0 • Both sequences converge. • lim n→∞ an = lim n→∞ bn Call this limit the arithmetic-geometric mean of a, b: M(a, b) = lim n→∞ an = lim n→∞ bn Why 1.19814023473559220744 etc is a Beautiful Number We have established that the arithmetic-geometric mean between 1 and √ 2 is π/ω to 11 places; the proof of this fact will certainly open up a new field of analysis. Gauss’s Diary, May 30, 1799 May 1800 - Gauss completes the proof Lemma 1. Let a, b > 0, a = a0, b = b0 and set I(a, b) = ∫ π/2 0 dq√ a2 cos2 q + b2 sin2 q . Then I(a, b) = I ( a+b 2 , √ ab ) May 1800 - Proof Complete! Lemma 1. Let a, b > 0, a = a0, b = b0 and set I(a, b) = ∫ π/2 0 dq√ a2 cos2 q + b2 sin2 q . Then I(a, b) = I ( a+b 2 , √ ab ) Lemma 2. I(a, b) = π/2 M(a, b) . May 1800 - Proof Complete! Lemma 1. Let a, b > 0, a = a0, b = b0 and set I(a, b) = ∫ π/2 0 dq√ a2 cos2 q + b2 sin2 q . Then I(a, b) = I ( a+b 2 , √ ab ) Lemma 2. I(a, b) = π/2 M(a, b) . Theorem. M(1 , √ 2) = πω. Further Reading J. M. Borwein and P. B. Borwein, Pi and the AGM (John Wiley & Sons, 1987) D. Bressoud A Radical Approach to Real Analysis (MAA, 1994). D. A. Cox, The arithmetic-geometric mean of Gauss, Enseignement Mathematique, 2nd ser., 30 (1984). 275-330. G. W. Dunnington, Gauss: Titan of Science, with additional material by Jeremy Gray (MAA, 2004). A. Rice. What Makes a Great Teacher? The Case of Augustus Morgan. The American Mathematical Monthly 106:6 (1999), 534-552.