Professional Documents
Culture Documents
JTG Problems 1 162 Convergence
JTG Problems 1 162 Convergence
1. Suppose we know that the volume of a pyramid with square base is a third of the product of
the height and the area of the base. We want to derive a formula for the volume of a
truncated square pyramid having lower base of side a, upper base of side b, and height h.
(a) If we put the top back onto the truncated pyramid so that the restored pyramid has height
ah
H, use similar triangles to show that H = .
a−b
(b) Now subtract the small upper pyramid from the large one to derive the formula
1
(
V = h a 2 + ab + b 2
3
)
for the volume of the truncated pyramid.
b Cross section
of “restored”
H
b b pyramid
h
a h
a a
Truncated pyramid
“If you are told: A truncated pyramid of 6 for the vertical height by 4 on the base by 2 on
the top, you are to square this 4, result 16; you are to double 4, result 8; you are to square
2, result 4; you are to add the 16, the 8, the 4, result 28; you are to take a third of 6, result
2; you are to take 28 twice, result 56. See, it is 56. You will find the formula right!”
Show that the formula in (b) yields the correct volume for this problem.
2. Derive the volume of the truncated pyramid in Problem #1(b) using the calculus technique
known as “volumes by slicing.”
__________________________________________________________________________________
William Dunham, “Problems for Journey Through Genius: The Great Theorems of Mathematics,” MAA
Convergence (July 2015), featuring “PROBLEMS for Great Theorems” © 1994 William Dunham
PROBLEMS for Great Theorems 2
IRRATIONALITY OF 2
One of the great discoveries of the School of Pythagoras was that the side and diagonal of a
square are incommensurable magnitudes – which, in numerical terms, says that 2 cannot be
written as the ratio of two integers. There follows a nice proof of the irrationality of 2 .
3. (a) If m is a positive integer, explain why each prime in the factorization of m2 must occur an
even number of times.
a
(b) Now suppose 2 is rational. That is, 2= for integers a and b. Square and cross-
b
multiply to conclude a2 = 2b2. From this, determine the number of times the prime “2”
occurs on each side of this equation and thereby derive a contradiction to the unique
factorization theorem.
4. Below is the figure Hippocrates used in his geometric proof of the quadrature of the lune. In
what follows, we re-prove his result algebraically (which is of course totally alien to the
thrust of Greek mathematics).
(a) Letting r be the radius of semicircle ACB, find Area (ΔAOC) in terms of r.
__________________________________________________________________________________
William Dunham, “Problems for Journey Through Genius: The Great Theorems of Mathematics,” MAA
Convergence (July 2015), featuring “PROBLEMS for Great Theorems” © 1994 William Dunham
PROBLEMS for Great Theorems 3
(b) Now determine the area of segment AFCD in terms of r. (Use the post-Hippocratean
fact that the area of a circle is π r 2 .)
(d) Using (b) and (c), find the area of lune AECF in terms of r.
(e) Finally, put all of this together to explain why Hippocrates’ lune is quadrable.
(a) Begin with a square of side 2r and upon each side construct a semicircle. Simultaneously
circumscribe a circle about the square (as shown). Find the combined area of the four
semicircles in terms of r.
__________________________________________________________________________________
William Dunham, “Problems for Journey Through Genius: The Great Theorems of Mathematics,” MAA
Convergence (July 2015), featuring “PROBLEMS for Great Theorems” © 1994 William Dunham
PROBLEMS for Great Theorems 4
(b) Now find the area of the circumscribed circle in terms of r. Notice anything?
(c) According to (a) and (b), we get the same area if we remove the circumscribed circle
from the figure above as we get if we instead remove the four semicircles from the figure
above.
Use this to explain why a single lune from the picture is quadrable.
6. Trace through the (fallacious) argument attributed to Hippocrates to the effect that the
quadrature of the lune implies the quadrature of the circle (Journey Through Genius, pp. 20-
22). Be sure you can identify where the error occurs. Does anyone seriously believe this
would have deceived the great Hippocrates?
p(x ) = an x n + an −1 x n −1 + ... + a2 x 2 + a1 x + a0 ,
where the coefficients are integers, such that p(c) = 0. Then consider the polynomial
q(x ) = an x kn + an −1 x kn − k + ... + a2 x 2 k + a1 x k + a0 .
__________________________________________________________________________________
William Dunham, “Problems for Journey Through Genius: The Great Theorems of Mathematics,” MAA
Convergence (July 2015), featuring “PROBLEMS for Great Theorems” © 1994 William Dunham
PROBLEMS for Great Theorems 5
1
9. Prove that, if c is algebraic, so is c . Of course, you’ll need to find a polynomial r(x) with
2
⎛ 1 ⎞
integer coefficients such that r ⎜ c ⎟ = 0 .
⎝ 2 ⎠
12. As we noted, in 1882 Ferdinand Lindemann proved that π was transcendental. Use the
previous problems to explain why π7, π + π, π − 1, and 1/π are transcendental as well.
14. (a) Do the proof of Proposition I.18. Note that it hinges on the Exterior Angle Theorem.
(b) Do the proof of Proposition I.19. This is a “double reductio ad absurdum” argument –
that is, it shows that of the three possibilities AC > AB, AC = AB, and AC < AB, two
lead to contradictions.
(c) Now prove Proposition I.20: “In any triangle, two sides taken together in any manner are
greater than the remaining one” – the Triangle Inequality the Epicureans thought was
patently obvious even to an ass.
15. None of the propositions in Book I prior to I.29 uses the Parallel Postulate in its proof,
whereas all of the later results in Book I depend on the Parallel Postulate, with a single
exception. Find it. Speculate as to why Euclid didn’t put it before I.29.
__________________________________________________________________________________
William Dunham, “Problems for Journey Through Genius: The Great Theorems of Mathematics,” MAA
Convergence (July 2015), featuring “PROBLEMS for Great Theorems” © 1994 William Dunham
PROBLEMS for Great Theorems 6
One thing that some people find surprising about Book I of the Elements is not the familiar
propositions that appear but the familiar ones that are omitted. Everyone may have his or her
own “favorite” that appears nowhere in Euclid, but the following are mine:
Through a point not on a line there can be drawn one and only one parallel to the given line.
Assume that you are inserting this in the Elements as Proposition I.31½.
From a point not on a line there can be drawn one and only one perpendicular to the given
line.
Note that proving the uniqueness of parallels (Problem #16) requires the Parallel Postulate,
but proving the uniqueness of perpendiculars (Problem #17) does not.
18. Prove that in a right triangle the midpoint of the hypotenuse is equidistant from the triangle’s
three vertices. Give a proof that could be inserted immediately after Proposition I.32. Be
careful – this one’s a bit tricky.
19. Now suppose Euclid wanted to follow up Problem #18 with a proof of the Side-Side
congruence scheme for right triangles – that is, if two right triangles have two sides of one
respectively congruent to two sides of the other, then the triangles are themselves congruent.
This is easy to do if we already have proved the Pythagorean Theorem, since that result
transforms Side-Side into Side-Side-Side. But suppose Euclid wanted to prove it prior to
Proposition I.47. Provide the proof. Note that there are really two cases – when the equal
sides are both legs (easy as π!) and when the equal sides are a leg and a hypotenuse.
Certainly no mathematical proposition boasts a greater number of different proofs than the
theorem of Pythagoras, for which well over 400 different arguments can be found in E.S.
Loomis’ The Pythagorean Proposition (NCTM, 1968). These, along with Euclid’s proof from
the Elements, may seem to approach mathematical overkill; on the other hand, they should
satisfy even the most hard-boiled skeptic.
For each proof outlined below, we begin with the right triangle BAC as shown below.
As you work through the details of the proofs in Problems #20–26, ask yourself where in each
one we use the assumption that we have a right angle.
__________________________________________________________________________________
William Dunham, “Problems for Journey Through Genius: The Great Theorems of Mathematics,” MAA
Convergence (July 2015), featuring “PROBLEMS for Great Theorems” © 1994 William Dunham
PROBLEMS for Great Theorems 7
b
c
β γ
B a C
20. The proof that many attribute to Pythagoras himself (i.e., the “ox-killer proof”):
c b b c
c c c
b
b b b
c b c b
(b) Prove that the inner figure on the right is itself a square.
__________________________________________________________________________________
William Dunham, “Problems for Journey Through Genius: The Great Theorems of Mathematics,” MAA
Convergence (July 2015), featuring “PROBLEMS for Great Theorems” © 1994 William Dunham
PROBLEMS for Great Theorems 8
21. This proof is due to the 12th century Hindu mathematician Bhaskara:
a
c b c
a b a
b
c b c
a
(c) Equate the area of the large square with the total areas of the small square and the four
triangles to prove the Pythagorean Theorem.
22. This proof of the Pythagorean Theorem is usually credited to the 17th century British
mathematician John Wallis, although it surely had been discovered prior to him. It is
regarded as the shortest proof of all.
(a) From A draw altitude AD to the hypotenuse and prove ∆ACD ~ ∆BAC ~ ∆BDA.
b a c a
(b) Conclude that = and = . From this, complete the proof.
CD b DB c
__________________________________________________________________________________
William Dunham, “Problems for Journey Through Genius: The Great Theorems of Mathematics,” MAA
Convergence (July 2015), featuring “PROBLEMS for Great Theorems” © 1994 William Dunham
PROBLEMS for Great Theorems 9
In right triangle BAC, mark off BD = BA, then bisect ∠ABC by line BE, where E is on side
AC. (By the way, could you give a convincing proof – independent of the diagram –
guaranteeing that the line bisecting ∠ABC MUST in fact meet AC?) Also, draw ED and call
its length x.
A
x
γ
B D C
(c) Set up the resulting proportions from (c), use these to eliminate x and thereby derive the
Pythagorean Theorem.
Begin with right ΔBAC and extend BC to L so that AC = LC. Construct CD bisecting
∠ACL, where D is on segment BA extended (as shown). Draw DL.
__________________________________________________________________________________
William Dunham, “Problems for Journey Through Genius: The Great Theorems of Mathematics,” MAA
Convergence (July 2015), featuring “PROBLEMS for Great Theorems” © 1994 William Dunham
PROBLEMS for Great Theorems 10
(d) Now attack the area of large ∆BLD in two ways: First, treat it as one large triangle;
second regard it as the amalgamation of the three sub-triangles. Equate the areas of these
figures, do a bit of algebra, and presto!
A
c b
β γ b
B a C L
__________________________________________________________________________________
William Dunham, “Problems for Journey Through Genius: The Great Theorems of Mathematics,” MAA
Convergence (July 2015), featuring “PROBLEMS for Great Theorems” © 1994 William Dunham
PROBLEMS for Great Theorems 11
bc
(d) Why is Area (ΔBAC) = ?
2
(e) Decomposing ΔBAC into the square and the two pairs of congruent triangles, prove that
(f) Finally, equate the two expressions for Area (ΔBAC) from (d) and (e), substitute for r
from (b), and do some algebra to derive the Pythagorean Theorem.
26. The last proof here is due to Congressman (later President) James A. Garfield of Ohio, who
published it in the New England Journal of Education in 1876.
(a) As a preliminary, prove that the area of a trapezoid is half the product of the height and
the sum of the bases.
γ
E
b a
c
β
A c B b D
__________________________________________________________________________________
William Dunham, “Problems for Journey Through Genius: The Great Theorems of Mathematics,” MAA
Convergence (July 2015), featuring “PROBLEMS for Great Theorems” © 1994 William Dunham
PROBLEMS for Great Theorems 12
(e) Find the area of the trapezoid in two different ways: first by using the formula from part
(a) and second by assembling the trapezoidal area as the sum of its three component
triangles. From this, prove the Pythagorean Theorem.
27. The previous seven proofs give an array of possible ways of establishing the theorem of
Pythagoras, but it should be noted that Euclid used none of them in the Elements. Which of
these could he have used as Proposition I.47 and which could he not have used? In this light,
do you give him high marks for the proof he actually devised for the Elements, or did he miss
an easy one?
28. Begin with ΔBAC and assume that a2 = b2 + c2. From B construct BC ⊥ AC and let AD have
length x.
A
x D
c b
B a C
(b) Now solve the equation from part (a) for x and go on to explain why Δ BAC must be
right. Where in here did you use the hypothesis?
29. The previous argument, like Euclid’s Proposition I.48, proves the converse of the
Pythagorean theorem by using the Pythagorean Theorem itself. But of course we do not
always use a theorem in the proof of its converse (e.g., Propositions I.27 and I.29). I’d be
curious (1) if anyone could find a proof of the converse of the Pythagorean Theorem that
does not require the Pythagorean Theorem itself. I’d be more curious (2) if someone could
find a proof of the converse that would fit perfectly as Proposition I.46½. And, I’d be most
curious (3) if someone, having proved the converse as Proposition I.46½, could then derive
the Pythagorean Theorem from the converse instead of vice versa. In short, could Euclid
have reversed the order of proof of these two great results while still keeping them at the end
of Book I?
__________________________________________________________________________________
William Dunham, “Problems for Journey Through Genius: The Great Theorems of Mathematics,” MAA
Convergence (July 2015), featuring “PROBLEMS for Great Theorems” © 1994 William Dunham
PROBLEMS for Great Theorems 13
The next problem – Problem 30 – is a proof of the converse independent of Proposition I.47
that satisfies (1), but it requires results about circles and similar triangles, which don’t come
until Books III and IV of the Elements. As to queries (2) and (3), I haven’t a clue.
30. The following proof of the converse uses circles and similarity.
With center B and radius BC, construct a circle (Postulate 3). Extend BA in both directions,
forming a diameter that meets the circle in D and E (Postulate 2). Next, extend CA to F on
the circle (Postulate 2) and draw CE, DF, and BF.
(e) Where in your argument did you use the hypothesis? (that a2 = b2 + c2)
a
b
c
D E
B A
__________________________________________________________________________________
William Dunham, “Problems for Journey Through Genius: The Great Theorems of Mathematics,” MAA
Convergence (July 2015), featuring “PROBLEMS for Great Theorems” © 1994 William Dunham
PROBLEMS for Great Theorems 14
For the next few problems, suppose we are operating under Saccheri’s Hypothesis of the Acute
Angle (HAA). That is, we suppose that the congruent summit angles of any Saccheri
Quadrilateral are acute. The thrust of these problems is to prove (1) that under HAA, triangles
have fewer than 180° and (2) that under HAA the Pythagorean relationship between hypotenuse
and legs of a right triangle cannot hold. This development is taken from Richard Trudeau’s The
Non-Euclidean Revolution.
31. Under HAA, prove that if ABCD is a Saccheri Quadrilateral as shown, then AB ≠ CD .
D C
A B
32. Again, we are working under HAA. Let ΔABC be given. Bisect AB at D and AC at E.
Draw line L between D and E and construct BF ⊥ L, CG ⊥ L, and AH ⊥ L.
C B
H L
G E D F
1
(c) Prove that DE = FG
2
__________________________________________________________________________________
William Dunham, “Problems for Journey Through Genius: The Great Theorems of Mathematics,” MAA
Convergence (July 2015), featuring “PROBLEMS for Great Theorems” © 1994 William Dunham
PROBLEMS for Great Theorems 15
(d) Finally, prove that the sum of the measures of the three angles in ΔABC is just
∠FBC + ∠GCB.
33. Now prove that under HAA, the angle sum of a triangle must be less than 180°.
34. Finally, we’ll prove that under HAA, the Pythagorean Theorem is not valid. Of course, we
could immediately reject the Pythagorean Theorem since “squares” do not exist under HAA
(that is, under HAA, all quadrilaterals contain fewer than 360° and thus there is no
quadrilateral possessing four right angles). But we could instead interpret the Pythagorean
Theorem as a relationship about (arithmetic) squares; while the “square of AB” might be
meaningless under HAA, “a2” is still a numerical value. So, the goal is to show that under
HAA it is false that a2 = b2 + c2 for right triangles.
C B
H L
G E D F
Assume, for the purpose of contradiction, that the Pythagorean Theorem holds. Begin with
right triangle BAC as shown above and repeat the constructions of Problem #32.
(b) Now derive a contradiction by looking back at Problem #31. Conclude that the
Pythagorean Theorem fails in hyperbolic geometry.
35. We showed that a valid congruence scheme in hyperbolic geometry is AAA. What about
Angle-Angle? That is, if two triangles have two angles of one congruent to two
corresponding angles of another, must the triangles be congruent? Prove your result.
__________________________________________________________________________________
William Dunham, “Problems for Journey Through Genius: The Great Theorems of Mathematics,” MAA
Convergence (July 2015), featuring “PROBLEMS for Great Theorems” © 1994 William Dunham
PROBLEMS for Great Theorems 16
36. Run through the easy proof of Proposition II.1. Translate this into a simple algebraic
identity.
37. Prove Proposition II.13. Do you recognize a famous result from trigonometry here?
38. Prove Proposition III.1, the construction and proof for finding the center of a given circle.
39. Euclid gave his clever construction of the regular pentagon in Proposition IV.11, but
following is the more familiar and more algebraic one:
1 x
F 36° J H
1
(b) Now, for the regular pentagon. Begin with a circle of unit radius with center O and
diameter AB as shown below. Bisect OB at C and construct radius OD ⊥ AB. With
center C and radius CD , draw an arc cutting AB at E. This yields the remarkable ΔEDO.
Prove the following three properties about this triangle:
__________________________________________________________________________________
William Dunham, “Problems for Journey Through Genius: The Great Theorems of Mathematics,” MAA
Convergence (July 2015), featuring “PROBLEMS for Great Theorems” © 1994 William Dunham
PROBLEMS for Great Theorems 17
40. If you haven’t done so for a while, get out a (big) piece of paper and construct a regular
pentagon with the technique from Problem #39.
41. If you haven’t done so for a while, use your construction in Problem #40 to construct a
regular pentadecagon (i.e., 15-gon) based on Euclid’s argument from Proposition IV.16.
42. Use the table of primes to count the number of primes in each batch of 250 consecutive
integers – i.e., 1 – 250, 251 – 500, ... up to (say) 2001 – 2250. Can you see why one might
suppose we’d eventually run out?
43. A famous old problem is to find a polynomial function f ( x ) so that, for each positive integer
n = 1, 2, 3, … , the result f ( n ) is a prime. Among other things, this would give an alternate
proof of the infinitude of primes, since there must be infinitely many different outputs among
the numbers f (1) , f ( 2 ) , f (3) , … .
(a) Prove this last statement. That is, assume an nth degree polynomial only generated
finitely many different values among f (1) , f ( 2 ) , … and derive a contradiction.
(b) A good candidate for f ( x ) is f ( x ) = x2 + x + 41. Find and factor the first three integers
which, when substituted into this polynomial, yield composites.
44. Explain how you could find 100 consecutive numbers, none of which is prime. How about a
billion consecutive non-primes? (HINT: Factorials!)
__________________________________________________________________________________
William Dunham, “Problems for Journey Through Genius: The Great Theorems of Mathematics,” MAA
Convergence (July 2015), featuring “PROBLEMS for Great Theorems” © 1994 William Dunham
PROBLEMS for Great Theorems 18
This is one of the most troubling unsolved problems in mathematics, for it is so ridiculously
simple to state: any even number greater than or equal to 4 can be written as the sum of two
primes. Alas, no one from Goldbach to Erdös has proved it.
45. (a) Verify the Goldbach Conjecture for 38, 538, and 1988.
46. Suppose that tomorrow someone proved the Goldbach Conjecture. Show that it would then
follow that any integer (even or odd) greater than 5 can be written as the sum of three primes.
DE POLIGNAC’S CONJECTURE
47. In 1848, De Polignac asserted: “Every odd number can be expressed as the sum of a power of
2 and a prime.”
(a) Verify this for odd numbers 3, 5, 7, 9, 11, 13, 15, 17, and 19.
48. Recall that if p and p + 2 are both prime, we call them “twin primes.”
(b) How many pairs of twin primess are there between 100 and 199? Between 1500 and
1599?
49. Clearly, any positive integer must assume one of the six forms
6k, 6k + 1, 6k + 2, 6k + 3, 6k + 4, or 6k + 5
(a) Use this to show that any prime greater than 3 is either one more or five more than a
multiple of 6.
__________________________________________________________________________________
William Dunham, “Problems for Journey Through Genius: The Great Theorems of Mathematics,” MAA
Convergence (July 2015), featuring “PROBLEMS for Great Theorems” © 1994 William Dunham
PROBLEMS for Great Theorems 19
(b) Now suppose we say that if p, p + 2, and p + 4 are all prime, we’ll call them “triplet
primes.” While it is unknown whether there are finitely or infinitely many pairs of prime
twins, we can determine whether the following conjecture is true or false:
From Euclid’s Proposition IX.20, we know we never run out of primes. In addition, since all
primes but 2 are odd, we can divide the odd primes into two categories – those that are one more
than a multiple of 4 (e.g., 5, 13, 17, 29, …) and those that are three more than a multiple of 4
(e.g., 3, 7, 11, 19, …). Obviously, at least one of these two categories of primes must be infinite.
In what follows, we’ll modify Euclid’s proof of Proposition IX.20 to show that there are
infinitely many primes of the form p = 4n + 3.
50. (a) Prove that the product of two numbers, each of which is one more than a multiple of 4, is
itself one more than a multiple of 4. In other words, if M = 4m + 1 and N = 4n + 1, then
MN also has this form. (Of course, all numbers in sight here are positive integers.)
(b) Now suppose that {a, b, c, d, …, e} is a finite collection of primes (in other words, an
“assigned multitude”), each having the form 4n + 3. We introduce the new number f =
4(abcd…e) − 1. Mimic Euclid’s two cases from Proposition IX.20 to show that there
must be a prime of the form 4n + 3 not among the original multitude. Conclude that there
are infinitely many primes of the form 4n + 3. (HINT: Unique Factorization Theorem
and part (a))
51. Are there finitely or infinitely many composites of the form 4n + 3? Explain.
52. One of the most remarkable theorems from number theory states that a prime of the form
4n + 1 can be written as the sum of the two perfect squares in one and only one way, whereas
a prime of the form 4n + 3 cannot be written as the sum of two perfect squares in any fashion
whatsoever. The first half of this theorem (originally stated by Fermat and first proved by
Euler) is difficult. The second half (i.e., the “4n + 3” case) isn’t too bad. Prove it.
(HINT: If p = 4n + 3 = a2 + b2, consider the three cases: a, b both even; a, b both odd; and
one of a, b even and the other odd.)
53. Show the Euler-Descartes formula holds for the Great Pyramid of Cheops, the World Trade
Center, and the Pentagon (which has a central courtyard and is thus not convex).
__________________________________________________________________________________
William Dunham, “Problems for Journey Through Genius: The Great Theorems of Mathematics,” MAA
Convergence (July 2015), featuring “PROBLEMS for Great Theorems” © 1994 William Dunham
PROBLEMS for Great Theorems 20
54. Now suppose we have a regular solid with V vertices, F faces, and E edges, where each face
of the solid is a regular n-gon. Suppose further that the number of regular n-gons meeting at
each vertex is k. Note immediately that n ≥ 3 and k ≥ 3.
(a) Initially, one might conclude that, since each of the F faces is a regular polygon with n
edges, there should be nF edges in the regular solid. Explain why the correct result is
E = nF/2.
(b) Explain why V = nF/k.
(c) Now substitute (a) and (b) into the Euler-Descartes formula to get
F
(2n + 2k − kn) = 1, and conclude that 2n + 2k − kn > 0.
4k
(d) The inequality in part (c) provides the key to the analysis of the regular solids. Begin with
the minimal case n = 3. Show that this implies that 3 ≤ k < 6 and thus yields three regular
solids.
(e) If n = 4, show the inequality yields only one regular solid.
(f) What happens if n = 5?, n = 6?, n = 7?, etc.
55. Complete the summary table below for the regular solids, combining the results of Euclid
with those from the previous page. You might look for interesting symmetries hidden among
the entries on the chart – you can be sure the Greeks would have.
1
Source: http://en.wikipedia.org/wiki/Platonic_solid
__________________________________________________________________________________
William Dunham, “Problems for Journey Through Genius: The Great Theorems of Mathematics,” MAA
Convergence (July 2015), featuring “PROBLEMS for Great Theorems” © 1994 William Dunham
PROBLEMS for Great Theorems 21
56. (a) Use calculus to verify that the area under the segment
of the parabola y = ax − x2 from x = 0 to x = a is 4/3
of the area of triangle OAB.
ARCHIMEDES’ APPROXIMATION OF π
t = 2 − 4 − s2
__________________________________________________________________________________
William Dunham, “Problems for Journey Through Genius: The Great Theorems of Mathematics,” MAA
Convergence (July 2015), featuring “PROBLEMS for Great Theorems” © 1994 William Dunham
PROBLEMS for Great Theorems 22
59. Now begin with a regular hexagon inscribed in a unit circle. The hexagon’s perimeter is 6, a
rough approximation for the circle’s circumference 2π, and so π ≈ 3.00. Now use Problem
#58 through seven doublings, until you have the perimeter of a regular inscribed 768-gon.
What is the corresponding approximation of π based on these inscribed figures?
60. In the midst of his approximation, Archimedes needed a value for 3 and he used
265 1351
< 3< . How good is this as a decimal?
153 780
61. An interesting question raised above is how one would get such a sharp approximation for
square roots without benefit of calculator. A very nice algorithm is the following:
(b) Now suppose we want to approximate 3 and we start with the rational number x0 = 5/3
(this is reasonable since (5/3)2 = 25/9 ≈ 27/9 = 3). Apply the recursion formula twice to
approximate this square root.
62. In his masterpiece On the Sphere and the Cylinder, Archimedes stated his results about
spherical volume and area by computing his figures with such better-understood figures as
cylinders and cones. Assuming we know the modern formulas for the key properties of
cones and cylinders, translate the following Archimedean statements into familiar, modern-
day formulas:
(a) “Any sphere is equal (by volume) to four times the cone which has its base equal to the
greatest circle in the sphere and its height equal to the radius of the sphere.”
(b) “Every cylinder whose base is the greatest circle in a sphere and whose height is equal to
the diameter of the sphere is half again as large as the sphere.”
__________________________________________________________________________________
William Dunham, “Problems for Journey Through Genius: The Great Theorems of Mathematics,” MAA
Convergence (July 2015), featuring “PROBLEMS for Great Theorems” © 1994 William Dunham
PROBLEMS for Great Theorems 23
(c) “Every cylinder whose base is the greatest circle in a sphere and whose height is equal to
the diameter of the sphere has surface (together with its bases) that is half again as large
as the surface of the sphere.”
63. Suppose we know that π = C/D is the constant ratio of a circle’s circumference to its
diameter. As noted, Proposition XII.2 of Euclid implies that, for circular area, A = k2D2 for
some constant k2, and Proposition XII.18 of Euclid yields that, for spherical volume,
V = k3D3 for some constant k2. But Euclid, of course, gave no clue as to the relationships
among the one-dimensional constant π, the two-dimensional constant k2, and the three-
dimensional constant k3. Use Archimedes’ insights from Measurement of a Circle and On
the Sphere and the Cylinder to express k2 and k3 in terms of π.
64. Recall that a spherical segment is that part of a sphere cut off by a plane. Archimedes, in
Proposition 42 of Book I of On the Sphere and the Cylinder, determined the surface area of a
segment (as we’ve seen). But he wasn’t finished. In Proposition 2 of the second book of On
the Sphere and the Cylinder, he polished off the volume of a segment as well. He proceeded
as follows:
x 2r + b
=
r −b r +b
1
Volseg = π (r − b )2 (2r + b ) .
3
__________________________________________________________________________________
William Dunham, “Problems for Journey Through Genius: The Great Theorems of Mathematics,” MAA
Convergence (July 2015), featuring “PROBLEMS for Great Theorems” © 1994 William Dunham
PROBLEMS for Great Theorems 24
(c) The only trouble with the formula from (a) is that it expressed the volume of the segment
in terms of r and b, which are parameters of the original sphere. However, if one were
presented with an actual segment, one would know only the dimensions h and c of the
segment itself, since the rest of the sphere would have been discarded. Thus, it makes
sense to adjust the formula of (a) to introduce h and c at the expense of r and b. Show
that such an adjustment yields:
1
Volseg = π h (3c 2 + h 2 ) .
6
(d) Finally, return to part (c) and determine what happens to the formula
ALGEBRAICALLY if h = r, while also determining what happens to
the segment GEOMETRICALLY if h = r. Put these together to re-
derive the formula for spherical volume.
65. A sector of a sphere is a region formed by intersecting the sphere with a cone having its
vertex at the sphere’s center. Thus, a sector looks sort of like an ice cream cone. In
Proposition 44 of Book I of On the Sphere and the Cylinder, Archimedes proved:
“The volume of any sector of a sphere is equal to a CONE whose base is equal to the surface
of the segment of the sphere included in the sector, and whose height is equal to the radius of
the sphere.”
In other words, he took the shaded area on the left – i.e., the area of the “surface of the
segment,” squashed it flat into a circle of equal area, and then upon the base erected a cone
with height equal to the sphere’s radius, as shown:
r Volume of Volume of
O sector = cone
Note the fascinating mixing of the segment’s surface – a surface in three dimensions – and
the equal plane area of the cone’s base.
__________________________________________________________________________________
William Dunham, “Problems for Journey Through Genius: The Great Theorems of Mathematics,” MAA
Convergence (July 2015), featuring “PROBLEMS for Great Theorems” © 1994 William Dunham
PROBLEMS for Great Theorems 25
(a) By adding the volume of the cone and the segment atop (from the previous problem),
show
π
Volsec =
6
( 2rc 2 + hc 2 + h 3 ) (same notation as in Problem #64)
(c) Finally, show that Archimedes’ peculiar statement with the squashed circle yields
precisely this same volume for a spherical sector.
66. We own a four-sided piece of land, whose area we must determine for tax purposes. Pacing
off the sides, we get lengths of 100 yards, 170 yards, 250 yards, and 240 yards (as shown),
and then we march along the diagonal (dotted) and find its length to be 260 yards. How many
square yards of property do we own?
A
67. Consider the triangle shown at the right.
__________________________________________________________________________________
William Dunham, “Problems for Journey Through Genius: The Great Theorems of Mathematics,” MAA
Convergence (July 2015), featuring “PROBLEMS for Great Theorems” © 1994 William Dunham
PROBLEMS for Great Theorems 26
69. Suppose we have right triangle BAC with right angle BAC and sides of length a, b, and c, as
shown. We extend CA to D so that AD = AC = b and then we draw BD.
B
a c
C D
b A
1
(d) Apply the simple formula for triangular area, Area = (base)(height), to prove that the
2
area of ΔDBC is bc.
(e) Finally, equate the area expressions from (c) and (d), simplify algebraically, and deduce
the theorem of Pythagoras.
70. (a) Beginning with a right triangle of sides a, b, and c, determine its area
1
in two ways – by the standard A = (base)(height) – and by Heron’s
2
formula. Equate these and manipulate the resulting, dreadful equation a
until you have derived the theorem of Pythagoras. b
(b) Which proof of Heron ⇒ Pythagoras do you like better, the one in
Problem #69 or the one in this problem? c
71. Do the arguments of Problem #69 or Problem #70 provide valid proofs of the Pythagorean
Theorem? That is, do they contain circular reasoning?
__________________________________________________________________________________
William Dunham, “Problems for Journey Through Genius: The Great Theorems of Mathematics,” MAA
Convergence (July 2015), featuring “PROBLEMS for Great Theorems” © 1994 William Dunham
PROBLEMS for Great Theorems 27
72. While I am not a fan of algebraic proofs of Heron’s Formula, the following is at least a little
better than most. It was sent to me by John C. Torrez of Chicago:
Begin with ΔABC having sides of lengths a, b, and c, as shown, where AC is at least as long
as the other two sides. Draw perpendiculars to AC at A and C and draw the line through B
parallel to AC and meeting the perpendiculars at E and D, respectively. Draw the altitude
from B to AC and call its length h. Also, let x = EB .
x B
E D
a
c h
A C
b
b2 + c 2 − a 2
(a) Using ΔAEB and ΔCDB, show x = .
2b
1
(b) Show Area (ΔABC) = K = b c 2 − x 2 .
2
(c) Using (a) and (b), derive the relatively simple expression:
1 2
(
K = b 4b 2c 2 − b 2 + c 2 − a 2 .
4
)
(Note that this is a difference of squares)
1
K=
4
((b + c) 2
− a 2 ) ( a 2 − (b − c)2 )
(e) Now factor the expression beneath the radical and derive Heron’s formula.
73. Here is a derivation of Heron’s formula, from the Law of Cosines. It has found its way into a
number of texts.
__________________________________________________________________________________
William Dunham, “Problems for Journey Through Genius: The Great Theorems of Mathematics,” MAA
Convergence (July 2015), featuring “PROBLEMS for Great Theorems” © 1994 William Dunham
PROBLEMS for Great Theorems 28
c a
α
A C
b
1
(a) Prove Area (ΔBAC) = bcsin α .
2
(b) Now begin with the Law of Cosines: a 2 = b 2 + c 2 − 2bc cos α . Add and subtract 2bc from
the right side and deduce that
(c) Again begin with the Law of Cosines, but this time alter your argument a bit to prove
(d) Finally, multiply the results of (b) and (c), mix in a pinch of (a), and end up with a proof
of Heron’s formula.
74. This is what I called an “Ancient/Modern” proof of Heron’s formula when I contributed it to
Mathematics Teacher in 1986.
Begin with two copies of ΔBAC. For the one on the left, mimic Heron’s argument by
inscribing a circle, generating x = s − b, y = s − c, and z = s − a, as shown. For the one on
the right, just draw the altitude from B to E, having length h. Also, let α = ∠BAC.
B B
x x
r c=x+z a=x+y
r
z y
O
r
α/2 α E
C A C
z D y b=y+z
A
__________________________________________________________________________________
William Dunham, “Problems for Journey Through Genius: The Great Theorems of Mathematics,” MAA
Convergence (July 2015), featuring “PROBLEMS for Great Theorems” © 1994 William Dunham
PROBLEMS for Great Theorems 29
h
(b) For the right-hand triangle, show: sin α = .
x+z
(e) Using the famous identity that sin ( 2θ ) = 2sin θ cosθ , with θ = α/2, apply (b), (c), and (d)
to prove
h 2rz
= 2 .
x + z r + z2
(f) Cross multiply and simplify (f) to show: s(r2 + z2) = z(xy + sz).
(g) Cross multiply and simplify (f) to show: s(r2 + z2) = z(xy + sz).
(i) Finally, combine (a) and (h) to show K2 = s(xyz) and from this deduce Heron’s formula.
75. Here is a wonderful trigonometric proof of Heron’s formula, sent to me by Barney Oliver,
who works at NASA at Moffett Field, California (don’t ask why he was working on Heron’s
formula instead of space flight). This one uses the trigonometric identities:
sin (α ± θ) = sin α cos θ ± cos α sin θ cos (α ± θ) = cos α cos θ sin α sin θ
B
x x
c
r a
z r
O y
r
α/2 γ/2
A D C
z y
b
__________________________________________________________________________________
William Dunham, “Problems for Journey Through Genius: The Great Theorems of Mathematics,” MAA
Convergence (July 2015), featuring “PROBLEMS for Great Theorems” © 1994 William Dunham
PROBLEMS for Great Theorems 30
tan δ + tan θ
(a) Prove that, for any angles δ and θ, that tan (δ + θ ) = .
1− tan δ tan θ
sin(δ + θ ) !
(HINT: tan(δ + θ ) = #
cos(δ + θ ) "
(b) Prove that if δ + θ, = 90°, then tan δ tan θ = 1 .
α β γ
(c) Show that if α, β, and γ are three angles of our triangle, then + + = 90°.
2 2 2
α !β γ $
(d) For α, β, and γ in our triangle, explain why tan
tan # + & = 1 .
2 " 2 2%
α β α γ β γ
(e) Now use (d) to prove that 1 = tan tan + tan tan + tan tan .
2 2 2 2 2 2
! β γ $$
(HINT: Somewhere in here you want to apply (a) to tan # + &&
" 2 2 %%
r2 r2 r2
(f) Next, use (e) to prove 1 = + + .
xz yz xy
xyz
(g) Get a common denominator and simplify (f) to conclude that r 2 = .
s
(h) Now use (g) to derive Heron’s formula.
76. Here we adapt the geometric/algebraic ideas of Cardano’s Ars Magna to an old friend – the
quadratic equation ax2 + bx + c = 0.
__________________________________________________________________________________
William Dunham, “Problems for Journey Through Genius: The Great Theorems of Mathematics,” MAA
Convergence (July 2015), featuring “PROBLEMS for Great Theorems” © 1994 William Dunham
PROBLEMS for Great Theorems 31
DEPRESSING POLYNOMIALS
77. (a) Prove by induction that if n ≥ 1, then the first two terms in the expansion of (y − r)n
are yn − nryn-1.
(b) Suppose we begin with the nth degree polynomial axn + bxn-1 + cxn-2 + … + dx + e.
b
Use (a) to verify that the substitution x = y − converts the nth degree polynomial
na
in x into an nth degree polynomial in y that lacks its (n−1)st degree term – i.e., one
that is depressed.
78. Use Cardano’s formula to get one real solution of x3 + 63x = 316.
79. Use Cardano’s formula to show that one real solution of the cubic x3 − 24x + 96 = 0 is
x = 23 2 7 −6 −23 2 7 +6
80. Use Cardano’s formula to get one real solution of 9x3 − 9x = 4 and retain all roots in order to
express your solution in terms of “radicals” (rather than using a decimal approximation).
Simplify your solution by radicals as far as you can.
5
81. (a) Find a “solution by radicals” of the depressed cubic x3 − 3x = .
2
(b) Use Cardano’s formula to find ALL real solutions to x3 − 6x2 + 9x − 4 = 0.
82. Use Cardano’s formula to get one real solution of the simple cubic ax3 + b = 0. Your result,
of course, should agree with what you can easily see to be the solution.
83. Cardano’s original example of his method in the Ars Magna was to solve the depressed cubic
x3 + 6x = 20.
(a) Show, as he did that the solution is x = 3 10 + 108 − 3 −10 + 108 .
(b) Cardano somehow recognized his solution as being simply “2.” Of course, he didn’t do
this on a calculator. Find an algebraic confirmation that x = 2.
(HINT: 10 + 108 = 10 + 6 3 = 1+ 3 3 + 9 + 3 3 = ( ? )3.)
__________________________________________________________________________________
William Dunham, “Problems for Journey Through Genius: The Great Theorems of Mathematics,” MAA
Convergence (July 2015), featuring “PROBLEMS for Great Theorems” © 1994 William Dunham
PROBLEMS for Great Theorems 32
84. In the late 1500s, Francois Viete devised an alternate approach to solving the cubic.
m
Beginning with the depressed cubic x3 + mx = n, he made the substitution x = − y.
3y
Show that this reduces the original cubic to a quadratic in y3 (which can thus be solved
for y3, then for y, then for x).
85. (a) Re-do Problem #78 with Viete’s method (your answers, of course, should agree).
It was Rafael Bombelli whose 1572 Algebra took the first successful steps toward resolving the
troubling “irreducible case of the cubic.” This case arose from the depressed cubic x3 + mx = n in
n 2 m3
the situation where + < 0 . Here Cardano’s formula introduces the square root of a
4 27
negative number. Bombelli realized that, in such a case, the real solutions of a real cubic require
us to detour into the realm of the complex numbers. That is, while we begin and end in the
comfortable domain of the reals, we seem to have to journey into the strange, uncharted world of
the imaginaries to complete the trip. To the sixteenth century mind, this seemed mighty strange.
(Come to think of it, it still does.)
86. (a) Recall that i2 = −1, i3 = − i, and i4 = 1. Use this to verify that (5 + i)3 = 110 + 74i.
Conclude that 3 110 + 74i = 5 + i .
(c) Use these cube roots and Cardano’s formula to find one real solution of x3 − 78x = 220.
3
(a) Verify that − 9 + 46i = 3 − 2i .
(b) Now find all real solutions for the real cubic x3 − 3x2 − 36x + 56 = 0 by first depressing
and then solving the resulting equation via Cardano. How many real solutions did you
get?
__________________________________________________________________________________
William Dunham, “Problems for Journey Through Genius: The Great Theorems of Mathematics,” MAA
Convergence (July 2015), featuring “PROBLEMS for Great Theorems” © 1994 William Dunham
PROBLEMS for Great Theorems 33
3
(a) Verify that 259 + 286i = 7 + 2i .
(b) Now find all real solutions for the real cubic x3 + 3x2 − 156x − 676 = 0. How many real
solutions did you get for this cubic?
NOTE: The previous three examples are artificial, since the cube roots of the pertinent complex
numbers seem to appear out of nowhere. There is a technique discovered by Euler in the
eighteenth century whereby the cube roots of any complex number can be found, although it
involves introducing the trig functions (!).
Not to be outdone, Viete likewise made an assault on the irreducible case. Interestingly, his
approach avoided the need for complex numbers; alas, it moved beyond the “algebraic” by
introducing the trigonometry of the reals. We present Viete’s idea in the following and then
try a pair of examples.
3 1
89. (a) Prove the trig identity cos3 α − cos α = cos3α .
4 4
(b) Now, to solve the depressed cubic x3 + mx = n, Viete made the substitution x = b cos α,
where α and b are unknowns to be determined. Use (a) to show
90. Apply Viete’s technique to get one real solution of x3 − 78x = 220. You may resort to
decimals and calculators here. Your result should, needless to say, agree with that obtained
in Problem #86(c).
91. Apply Viete’s technique to get one real solution of x3 − 6x = 4. This time, express your
answer as a “solution by radicals” if you can – no decimal approximation and no trig
functions in your final answer!
(HINT: Along the way, you may want to derive the not-so-famous fact that
!
π 1 3#
cos = +
12 2 4 #"
__________________________________________________________________________________
William Dunham, “Problems for Journey Through Genius: The Great Theorems of Mathematics,” MAA
Convergence (July 2015), featuring “PROBLEMS for Great Theorems” © 1994 William Dunham
PROBLEMS for Great Theorems 34
92. Write out the first five terms (i.e., through x4) in the binomial expansion of 3 1+ x .
1
Use the previous problem to get a quick numerical estimate of ∫ 2 3
1+ x dx ,
0
94. Now CUBE the expansion you found in Problem #92, carrying the result at least through
the x3 term. What do you notice (besides hand cramps)?
AN ANALYTIC APPROXIMATION OF π
x3 x5 x 7
Here is an informal derivation of the series: tan −1 x = x − + − + ...
3 5 7
x 1
(a) Use the substitution t = tan α to show ∫ 2
dt = tan −1 x .
0
1+ t
1 −1
(b) Now expand 2
= (1+ t 2 ) by the binomial series and integrate the series term-wise
1+ t
from 0 to x, and thereby derive the Gregory series.
96. Using trig identities (NO CALCULATORS ALLOWED!!), prove the peculiar formula of
John Machin (1706):
π ⎛ 1 ⎞ ⎛ 1 ⎞
= 4 tan −1 ⎜ ⎟ − tan −1 ⎜ ⎟
4 ⎝ 5 ⎠ ⎝ 239 ⎠
"1% " 1 %
(HINT: Introduce α = tan −1 $ ' and β = tan −1 $ ' and then attack β = tan ( 4α − β ) with
# 5& # 239 &
your trig identities.)
__________________________________________________________________________________
William Dunham, “Problems for Journey Through Genius: The Great Theorems of Mathematics,” MAA
Convergence (July 2015), featuring “PROBLEMS for Great Theorems” © 1994 William Dunham
PROBLEMS for Great Theorems 35
⎛ 1 ⎞
97. Now use the first four terms of the Gregory series to estimate both tan −1 ⎜ ⎟ and
⎝ 5 ⎠
⎛ 1 ⎞
tan −1 ⎜ ⎟ and then, from Machin’s identity, estimate π.
⎝ 239 ⎠
98. Finally, compare the efficiencies of the classical approach of Archimedes (see Problems
#58–59), of Newton, and of Machin.
(a) Which estimate gives the best accuracy for the buck?
(b) How many root extractions does the classical method need? Newton? Machin?
99. (a) Use your calculator or a computer to determine exactly how many terms we must sum
in the harmonic series before it totals 4.50 or more.
This constant is approximately equal to 0.577215. (It remains one of the great unsolved
problems of mathematics to determine whether the Euler-Mascheroni constant is or is not
rational.) In any event, use this value of γ to estimate the number of terms it takes the
harmonic series to grow to 10.00 or more.
(c) How long does it take the harmonic series to reach 15.00 ?
n
1 n
100. Prove by induction that ∑ k (k +1) = n +1 and use this to give a modern proof
k=1
∞
1
(i.e., one using partial sums) that ∑ k (k +1) = 1 .
k=1
101. We’ve seen Leibniz cleverly sum the series of reciprocals of triangular numbers:
1 1 1 1
1+ + + +... + +... = 2 .
3 6 10 k ( k +1) 2
__________________________________________________________________________________
William Dunham, “Problems for Journey Through Genius: The Great Theorems of Mathematics,” MAA
Convergence (July 2015), featuring “PROBLEMS for Great Theorems” © 1994 William Dunham
PROBLEMS for Great Theorems 36
But he didn’t stop there, for he next summed the reciprocals of the so-called “pyramidal
numbers”:
1 1 1 1 1
P = 1+ + + + +... + +... .
4 10 20 35 k ( k +1) ( k + 2 ) 6
2 2 2 2 2 2
(a) By considering P= + + + + +... and cleverly decomposing the
3 3 12 30 60 105
(b) Now re-do the sum by the modern technique of partial sums. That is, first prove
by induction that
n
1 3n ( n + 3)
∑ k (k +1) (k + 2) = .
k=1 6 2 ( n +1) ( n + 2 )
102. Here we follow a similar line of reasoning to sum the infinite series:
1 1 1 1 1 1 1
+ + + + + + ... + 2 + ...
3 15 35 63 99 143 4k − 1
(b) As Leibniz did with Huygens’ challenge problem, decompose the series 2S into
cancelling “+” and “−” terms, and thereby evaluate 2S and thus S.
2 1 2 1
(HINT: = 1 − ; = − ?; etc.)
3 3 15 3
(b) Can you find a specific geometric series which sums to −30?
__________________________________________________________________________________
William Dunham, “Problems for Journey Through Genius: The Great Theorems of Mathematics,” MAA
Convergence (July 2015), featuring “PROBLEMS for Great Theorems” © 1994 William Dunham
PROBLEMS for Great Theorems 37
1
(c) Can you find a specific geometric series which sums to − ?
3
(d) Determine precisely which real numbers can be sums of geometric series.
104. Perhaps to show up his brother Johann, Jakob Bernoulli devised his own clever proof for the
divergence of the harmonic series in 1689 and included it in his Tractatus de Seriebus
1
Infinitis. His key idea was to prove that, starting at any point of the harmonic series, the
a
1 1
sum + +... will, after a FINITE number of terms, exceed 1. In Jakob’s words, “No
a a +1
matter where we start in the harmonic series, the sum of a finite number of consecutive
terms will exceed 1.” Below appears a (streamlined) proof:
1 1 1
+ +... + 2 .
a +1 a + 2 a
1 1 1 1 1 1 1
(b) Explain why + +... + 2 ≥ 2 + 2 + 2 +... + 2
a +1 a + 2 a a a a a
1
= "#a 2 − a$% 2 .
a
(c) Now use (b) to show that, as Jakob claimed, the portion of the harmonic series
1 1 1
+ +... + 2 ≥ 1 .
a a +1 a
1 1 1 1 1 1
+ + ... + ≥ 1 and + + ... + ≥ 1 .
3 4 9 4 5 16
(e) Finally, explain why Jacob’s conclusion in (c) proves the harmonic series
diverges to +∞. What is so important about the finite number of terms?
__________________________________________________________________________________
William Dunham, “Problems for Journey Through Genius: The Great Theorems of Mathematics,” MAA
Convergence (July 2015), featuring “PROBLEMS for Great Theorems” © 1994 William Dunham
PROBLEMS for Great Theorems 38
∞
1
105. We noted that Jakob Bernoulli knew ∑k 2
converged to something less than 2 by
k=1
comparing this series with a term-wise greater series:
1 1 1 1 ⎡ 1 1 1 1 1 ⎤
1 + + + + + ... = 2⎢ + + + + + ...⎥ = 2(1) = 2 .
3 6 10 15 ⎣ 2 6 12 20 30 ⎦
2 2 2 2 2
+ + + + + ... = 1 .
3 15 35 63 99
1 1 1 1 1 1 1 1
(b) Now if we let S1 = + + + + ... , it is clear that S1 < + + + + ... .
4 16 36 64 3 15 35 63
1
Use this and part (a) to get the bound S1 < .
2
1 1 1 5
(c) Similarly, if S 2 = 1 + + + + ... , show S2 < .
9 25 49 4
∞
1
(d) Finally, explain why ∑k 2
< 1.75
k=1
106. Consider the example l’Hospital gave as the first (!) illustration of his rule in his 1696
Analyse de Infiniment Petit:
2a 3 x − x 4 − a 3 a 2 x
lim .
x→a
a − 4 ax 3
__________________________________________________________________________________
William Dunham, “Problems for Journey Through Genius: The Great Theorems of Mathematics,” MAA
Convergence (July 2015), featuring “PROBLEMS for Great Theorems” © 1994 William Dunham
PROBLEMS for Great Theorems 39
1 1 1 1 π2
1+ + + + +... = .
9 25 49 81 8
108. Now use the previous result to get an independent derivation of our great theorem:
1 1 1 1 1 π2
1+ + + + + +... = .
4 9 16 25 36 6
1 1 1 1
109. (a) Evaluate + + + +...
4 16 36 64
1 1 1 1
(b) Now evaluate the alternating series 1− + − + −...
4 9 16 25
110. Finally, use the product expansion for cos x from Problem #107 and a clever choice for x to
show
∞
1
EULER’S EVALUATION OF ∑k 6
k=1
111. Push Euler’s argument one step further to get a formula relating ∑r , ∑r
k k
2
, and ∑r k
3
1 1 1 1
1+ + + +... + 6 +...
64 729 4096 k
__________________________________________________________________________________
William Dunham, “Problems for Journey Through Genius: The Great Theorems of Mathematics,” MAA
Convergence (July 2015), featuring “PROBLEMS for Great Theorems” © 1994 William Dunham
PROBLEMS for Great Theorems 40
(a) a = 12 and p = 7.
114. (a) Is the Little Fermat Theorem still true if we drop the restriction that p not be a factor
of a (keeping everything else the same)?
(b) Is the Little Fermat Theorem still true if we drop the restriction that p be prime
(keeping everything else the same)?
116. CONJECTURE: “If a is any positive integer, then 7 must divide evenly into one of the
following: a, a + 1, a + 6, or a4 + a2 + 1.” Either prove this conjecture or find a
counterexample. HINT: The Little Fermat that could!)
117. Euler showed that a prime p divides evenly into a4 + 1 (a even) only if p = 8k + 1. Use this
insight to factor 234,257 into the product of primes. In so doing, keep track of how many
divisions you actually must do before hitting on a prime factor.
119. Recall we mentioned in Problem #52 that Euler proved Fermat’s assertion that a prime of
the form p = 4k + 1 can be written as the sum of two squares in one and only one way.
(a) Verify this fact numerically for the prime p = 4(22) + 1 = 89.
(c) Now use Euler’s result to explain why 232 + 1 must be composite.
(HINT: 232 + 1 = 418161601 + ? )
__________________________________________________________________________________
William Dunham, “Problems for Journey Through Genius: The Great Theorems of Mathematics,” MAA
Convergence (July 2015), featuring “PROBLEMS for Great Theorems” © 1994 William Dunham
PROBLEMS for Great Theorems 41
REPRESENTATION OF INTEGERS
120. Check by listing all possibilities that the number of ways of writing 12 as the sum of
distinct integers is equal to the number of ways of writing 12 as the sum of (not necessarily
distinct) odd integers.
121. We’ve seen that, in the eighteenth century, the Fundamental Theorem of Algebra was
framed in terms of factoring real polynomials into the product of real linear and/or real
quadratic factors. Recall that Euler managed to factor Nicolaus Bernoulli’s proposed
counterexample x4 − 4x3 + 2x2 + 4x + 4 into the product of
( ) (
x2 − 2 + 4 + 2 7 x + 1+ 4 + 2 7 + 7 and )
x − (2 −
2
4+2 7 ) x + (1− 4+2 7 + 7) .
122. Factor the real depressed quartic x4 + x2 + 4 into the product of two real quadratic factors
AND check your result. (HINT: Since this is depressed, if one factor looks like x2 + ux + a,
the other must look like x2 − ux + a.)
125. Now suppose that P(z) = anzn + an-1zn-1 + … + a2z2 + a1z + a0 is an nth degree polynomial for
which the coefficients an, an 1, … a1, a0 are all REAL numbers. Suppose also that z = z0 is a
−
complex solution of the equation P(z) = 0. Prove that z = z0 is also a solution of P(z) = 0. In
other words, as Euler observed in 1742, solutions to polynomial equations with real
coefficients come in conjugate pairs. (HINT: You might want to use the two previous
problems here.)
__________________________________________________________________________________
William Dunham, “Problems for Journey Through Genius: The Great Theorems of Mathematics,” MAA
Convergence (July 2015), featuring “PROBLEMS for Great Theorems” © 1994 William Dunham
PROBLEMS for Great Theorems 42
128. Now prove (cos θ − i sin θ)n = cos nθ − i sin nθ. (HINT: Use −θ in Problem #127.)
15 8
129. Use DeMoivre’s Theorem to get easy evaluations of ( 3 +i ) and of (3+ 4 i) .
130. We’ve seen that DeMoivre’s Theorem provides the vehicle for determining roots of any
complex numbers. Use this technique to find all three cube roots of 110 + 74i. Compare
this with the result of Problem #86. Notice how this technique allows us to overcome the
“irreducible case” of the cubic that so badly troubled Cardano.
131. We know that, for positive real numbers, square roots come in “± pairs” – e.g., the two
square roots of 9 are +3 and −3. Does this same principle hold for general complex
numbers? That is, if we apply our technique to find the two square roots of a complex
number z, and if one of these square roots is w, must the other be −w? Explain, using the
square root formula.
Then expand the binomial, equate the real and imaginary parts, and solve the resulting pair
of equations for a and b. (BEWARE: a and b are real numbers here.)
133. The algebraic technique of the previous problem seems, at first glance, to be a way to make
an end-run around DeMoivre when taking roots of complex numbers, and thus a purely
algebraic way to overcome the irreducible case of the cubic. However, before getting too
elated about all this, try to adapt this technique to find 65 + 142 i . That is, let
65 + 142 i = a + bi, so that (a + bi)3 = 65 +142i . Then expand, generate a pair of
equations in a and b, and boogie. (WARNING!! – This technique may be hazardous to
your health since it doesn’t work. Don’t spend too much time on this; just convince
yourself that some of life’s promising things somehow just don’t pan out.)
__________________________________________________________________________________
William Dunham, “Problems for Journey Through Genius: The Great Theorems of Mathematics,” MAA
Convergence (July 2015), featuring “PROBLEMS for Great Theorems” © 1994 William Dunham
PROBLEMS for Great Theorems 43
134. (a) Find all six 6th roots of unity and write them in rectangular form.
137. Factor the real cyclotomic polynomial x23 + x22 + … + x3 + x2 + x + 1 into the product of
real linear and/or real quadratic factors. Stop at any point when you want to wring my
neck. (HINT: π/12 = π/3 − π/4.)
138. Prove the peculiar trigonometric fact that for any positive integer n, we have:
1+ cos
2π
+ cos
4π
+ cos
6π
+... + cos
(2n − 2) π = 0 .
n n n n
sin
π
+ sin
2π
+ sin
3π
+... + sin
(n −1) π = cot π .
n n n n 2n
(a) First, derive the finite geometric series formula for any complex number z ≠ 1:
2 n-1 1− z n
1+z+z +…+z = .
1− z
π π
(b) Now let z = cos + i sin on both sides of the equation in (a); use DeMoivre,
n n
conjugates, trig, and whatever else to prove the identity. (Pretty neat, huh?)
140. Be sure you can show that, if a and b are constructible magnitudes, then so are the
magnitudes a + b, a − b, ab, a/b, and a (provided these are all real numbers).
141. For the 17-gon construction, explicitly show that a1a2 = −4 . (UGH!!)
__________________________________________________________________________________
William Dunham, “Problems for Journey Through Genius: The Great Theorems of Mathematics,” MAA
Convergence (July 2015), featuring “PROBLEMS for Great Theorems” © 1994 William Dunham
PROBLEMS for Great Theorems 44
142. For the 17-gon construction, find the explicit formula for b1 in terms of square roots; in
other words, show that b1 is indeed a surd. Of course, you’ll first have to determine a1 as a
surd. Good luck.
143. List all the values of n less than 200 for which a regular n-gon can be constructed with
compass and straightedge. (HINT: There are more than 30 of them.)
Recall the proof by contradiction went as follows: It began by assuming that, although this
cubic had no rational solution, it nonetheless did have a constructible solution of the form
r = p + q β , where this solution was “minimal” in the sense that it used as few or fewer
square roots in its formula than any other constructible solutions. He next showed that
s = p − q β will also be a root of the cubic above.
(a) Explain why this implies that a1a2 = −4 is also a root of the cubic.
(c) Now, carefully indicate exactly where in this argument we used the fact that the cubic
had no rational roots. That is, show precisely how the desired contradiction falls apart
if rational roots are permitted.
145. With the key result behind him, Wantzel proved that the 60° angle could not be trisected
with compass and straightedge. Use this fact to prove:
__________________________________________________________________________________
William Dunham, “Problems for Journey Through Genius: The Great Theorems of Mathematics,” MAA
Convergence (July 2015), featuring “PROBLEMS for Great Theorems” © 1994 William Dunham
PROBLEMS for Great Theorems 45
ONE-TO-ONE CORRESPONDENCES
πn
sin 2 + n cos π n
f (n) = 2
2
is our one-to-one correspondence between the set of natural numbers and the set of
(positive, negative, and zero) integers.
147. Find an explicit function that generates the same correspondence as that of Problem #147
but which has only algebraic – as opposed to trigonometric – components.
148. Write down an explicit function g : (0, 1) → [0, 1] that is both one-to-one and onto. Note
that this establishes [ 0,1] = ( 0,1) = c directly from the definition.
149. By finding an explicit one-to-one correspondence, prove that for any real number a, the
cardinality of the ray (a,+∞) is c.
150. What is the cardinality of the set of all complex numbers? Why?
151. In his first (1874) proof of the non-denumerability of the continuum, Cantor needed the
theorem from analysis which says that any nested sequence of closed intervals has at least
one point belonging to all the intervals. Today this is often called “Cantor’s Nested Set
Theorem.” Here we want to show why the nested intervals must be closed.
⎛ 1 ⎞ ⎛ 1 ⎞ ⎛ 1 ⎞
(0,1) ⊃ ⎜ 0, ⎟ ⊃ ⎜ 0, ⎟ ⊃ … ⊃ ⎜ 0, ⎟ ⊃ …
⎝ 2 ⎠ ⎝ 3 ⎠ ⎝ n ⎠
And let A be the set of points that belong to all of these intervals. Explain why there can be
nothing in A.
__________________________________________________________________________________
William Dunham, “Problems for Journey Through Genius: The Great Theorems of Mathematics,” MAA
Convergence (July 2015), featuring “PROBLEMS for Great Theorems” © 1994 William Dunham
PROBLEMS for Great Theorems 46
152. Let B be the set of all real numbers in the interval (0,1) whose decimal expansions contain
only the digits “0” and “5.” For instance, B contains the numbers 0.05, ½, and
0.05005000500005…, whereas 0.050050504555… is not a member of B. Use the
diagonalization process to explain why B is a non-denumerable subset of (0,1).
1
153. (a) Explain why ½ belongs to set B in Problem #152, and why also belongs to B.
20
More generally, list an infinite number of rationals belonging to B.
(b) Does set B from Problem #152 contain finitely or infinitely many irrationals? Use a
set theoretic argument to answer this question.
154. The height h of a polynomial with integer coefficients is the largest absolute value of any
coefficient of the polynomial. List all polynomials of height h = 4 and solve them to
determine which new algebraic numbers they yield – i.e., which new algebraic numbers
appear here that did not appear for heights h = 1, 2, or 3. (HINT: There are a dozen new
ones.)
π
155. Prove that cos is algebraic by exhibiting a specific polynomial with integer coefficients
7
π
for which cos is a zero. (HINT: 7α = 3α + 4α.)
7
156. (a) Show that P(x) = x2 − 3 and Q(x) = 4x3 − 8x − 1 are two different polynomials with
integer coefficients that take the same value for x = 2 .
(b) Using Lindemann’s 1882 result that π is transcendental, explain why there do NOT
exist two different polynomials P(x) and Q(x), each having integer coefficients, for
which P(π) = Q(π). Put another way, this says that, if for some polynomial P(x) with
integer coefficients, you find P(π) = a, then there is no other such polynomial Q(x)
taking the value a when x = π.
(c) Does any polynomial with integer coefficients have x-intercept at x = π? Does any
have y-intercept at y = π? (Which of these two parts is very elementary and which is
very deep?)
__________________________________________________________________________________
William Dunham, “Problems for Journey Through Genius: The Great Theorems of Mathematics,” MAA
Convergence (July 2015), featuring “PROBLEMS for Great Theorems” © 1994 William Dunham
PROBLEMS for Great Theorems 47
SCHRÖDER-BERNSTEIN IN ACTION
157. (a) Use Schröder-Bernstein to prove [0,1] = c (compare with Problem #148).
158. (a) Look over the proof in Journey Through Genius (pp. 270-271) that uses Schröder-
Bernstein to prove that the cardinality of the irrationals is c.
(b) Can you cook up a similar argument to establish that the cardinality of the set of
transcendentals is c? (I can’t – I need the continuum hypothesis for this, and that’s
quite a different thing.)
159. Determine the cardinality of the set of all constructible numbers – i.e., the set of real
numbers we called the “surds.” (HINT: This is still the section on Schröder-Bernstein,
after all.)
160. Let N = {1, 2, 3, …} be the set of natural numbers and O = {1, 3, 5, …} be the set of odd
natural numbers. Introduce the mappings
n+5
g : O → N defined by g(n) = .
2
(a) Show that f and g are both one-to-one but NOT onto.
(b) Use these specific functions to determine the sets E0, E1, E2, and E3 from the Schröder-
Bernstein proof, and thereby write down the first few members of the set
E = E 0 ∪ E1 ∪ E2 ∪ …
(c) Next, construct the specific formula for the one-to-one and onto function
h : N → O and give your answer in “piecewise” form:
? if n is in E
h(n) =
? if n is in N \ E.
(d) Finally, “check” your function h in (c) by making a chart showing h(1) , h(2) ,
h(3) , …, h(15) and noting whether h really does seem to be one-to-one and onto,
as Schröder-Bernstein promised.
__________________________________________________________________________________
William Dunham, “Problems for Journey Through Genius: The Great Theorems of Mathematics,” MAA
Convergence (July 2015), featuring “PROBLEMS for Great Theorems” © 1994 William Dunham
PROBLEMS for Great Theorems 48
CANTOR’S THEOREM
161. Prove by induction on n: If a set A has n elements, its power set has 2 n elements. (This,
of course, gives the finite version of Cantor’s Theorem.)
162. In his 1891 proof that the cardinality of the power set exceeds the cardinality of the set
itself, Cantor began with any matching between elements of the set A and elements of its
power set and had to come up with an element of the power set that could not be on the
proposed list. The fact that he figured out how to find such a set in general is a clear
testimony to his remarkable genius. Just to see how his insight works in a specific case or
two, work on the examples below:
(a) Suppose we match all of A = {a, b, c, d, e, f, g} with some of the elements of its power
set P(A) as follows:
a → {b, c, g}
b → {b, c, e, g}
c → {b, c}
d → {d}
e → A
f → {a, b, c}
g → { } (empty set)
Find the set B from Cantor’s proof. Is it an element of P(A)? Is it matched with
anything in the matching above?
(b) Do the same thing for the matching between N and its power set P(N) given by
n → {n2, n2 + 2, n2 + 4, …} if n is odd.
For this matching, explicitly find the subset B of N and explain why it is or isn’t
matched with any element n under this matching.
__________________________________________________________________________________
William Dunham, “Problems for Journey Through Genius: The Great Theorems of Mathematics,” MAA
Convergence (July 2015), featuring “PROBLEMS for Great Theorems” © 1994 William Dunham