Math 453, Section X13 Final Exam Solutions Spring 2011

Problem 1
True/false questions. For each of the following statements, determine if it is true or false, and provide
a brief justification for your claim. Credit on these questions is based on your justification. A simple
true/false answer, without justification, or with an incorrect justification, won’t earn credit.
A justification typically consists of citing and applying an appropriate theorem (e.g., “by Chamberlain’s
Theorem”), and if necessary stating why the cited theorem can be applied. Be specific; e.g., say “Since
(453, 347) = 1, ’s Theorem with a = 453 and b = 347 applies and guarantees the existence of a solution ...”
rather than something like “true by Euler’s Theorem”.

(i) An integer n > 2 is composite if and only if n does not divide (n − 1)! + 1.

Solution: T RU E Wilson’s theorem states that an integer n > 2 is prime if and only if (∗) (n−1)! ≡
−1 mod n. The given condition, “n does not divide (n − 1)! + 1”, is the negation of (∗), so it holds if
and only if n is composite.

(ii) There exists an integer x such that 3x ≡ 347 mod 453. (Note that 3 and 347 are prime, and 453 has
prime factorization 453 = 3 · 151.)

Solution: F ALSE A congruence ax ≡ b mod m has a solution if and only if (a, m) | b. In the given
case (a, m) = (3, 453) = 3, and since 3 - 347, no solution exists.

(iii) If n is an integer ≥ 2 satisfying n | an−1 − 1 for all positive integers a with (a, n) = 1, then n is prime.

Solution: F ALSE Composite integers n that satisfy the given condition, i.e., an−1 ≡ 1 mod n for all
a with (a, n) = 1, are Carmichael numbers, and these are known to exist (e.g., n = 561).
Remark: Note that the existence of pseudoprimes, i.e., composite numbers satisfying an−1 ≡ 1 mod n
for a given base a, is not enough to disprove the statement, since pseudoprimes are defined with respect
to a fixed base a. A Carmichael number must satisfy the above congruence for all bases a coprime with
n, which is a much more restrictive condition on n.

Problem 2
Definitions and theorems. The following problems test your knowledge of theorems and definitions: Simply
state the theorem or definition requested; be sure to include any necessary hypotheses, and be careful with
details (e.g., “for all a, b ∈ Z” versus “for all a, b ∈ N” or for all “a, b ∈ N such that (a, b) = 1”).
pi
(i) Give a precise statement of the “Best Approximation Property” of the convergents of the continued
qi
fraction expansion of a real number α.
Solution: For any rational number a/b, with a ∈ Z, b ∈ N, and b ≤ qi , we have |α − pi /qi | ≤ |α − r/s|.

(ii) State the Prime Number Theorem (be sure to define/explain any notation involved).
π(x)
Solution: The Prime Number Theorem asserts that limx→∞ = 1, where π(x) denotes the
x/ ln x
number of primes ≤ x.

(iii) State Dirichlet’s Theorem for primes in arithmetic progressions.

Solution: Let a, b ∈ N such that (a, b) = 1. Then there exist infinitely many primes of the form an + b,
where n ∈ N.

Page 1
Math 453, Section X13 Final Exam Solutions Spring 2011

Problem 3
Short computations. Each of the following questions can be answered using only a minimal amount
of pencil/paper calculations if approached with the right method. If you get entangled in a messy hand
computation (e.g., multiplying or dividing multidigit numbers), you are on the wrong track. Answers arrived
at by brute force methods, trial and error, or guessing won’t earn credit.
Make sure to show your work, cite any theorems you use (e.g., by “Dirichlet’s Theorem”), and circle/box
your final answer.
(i) Find the remainder of 72012 upon division by 2011.

Solution: Since 2011 is prime, Fermat’s Theorem gives 72010 ≡ 1 mod 2011, so 72012 ≡ 72 = 49 mod
2011

(ii) Find the last two decimal digits of 413402 . (Don’t try this with brute force. With the right approach,
this requires only minimal computations.)
Solution: The last two decimal digits of a positive integer n are given by the least nonnegative residue
of n modulo 100. Since 413 ≡ 13 mod 100, it suffices to compute 13402 mod 100. Euler’s theorem gives
13ϕ(100) = 1340 ≡ 1 mod 100 (note ϕ(100) = ϕ(22 · 52 ) = (2 − 1) · 21 · (5 − 1) · 51 = 40). Then

13402 = 1340·10+2 = (1340 )10 · 132 ≡ 110 · 132 ≡ 169 ≡ 69 mod 100.

(iii) Find all integer solutions (x, y) of the equation 13x + 11y = 7.
Solution: We first apply the Euclidean algorithm to 13 and 11 to obtain a single solution (x0 , y0 ) to
the equation 13x + 11y = 1:

13 = 1 · 11 + 2, 11 = 5 · 2 + 1, 5 = 5 · 1.

Reversing the steps gives

1 = 1 · 11 − 5 · 2 = 1 · 11 − 5 · (13 − 11) = 6 · 11 − 5 · 13,

This shows that (x0 , y0 ) = (−5, 6) is a solution to the equation 13x + 11y = 1.
Multiplying through by 7 gives a particular solution to 13x + 11y = 7: (x1 , y1 ) = 7(x0 , y0 ) = (−35, 42).
To get all solutions, we add an integer multiple of 11 to x0 and subtract the same multiple of 13 from
y0 : (x, y) = (−35 + 11k, 42 − 13k), k ∈ Z

Problem 4
Short proofs.
(i) Prove that if 2n − 1 is a prime number, then n is also a prime number.
Solution: We show the equivalent statement that if n is composite, then 2n − 1 is also composite.
Suppose n is composite. Then n = ab for some integers a, b ≥ 2. Since 2a ≡ 1 mod (2a − 1), we have
2n = (2a )b ≡ 1b = 1 mod (2a − 1). Thus, 2n − 1 is divisible by 2a − 1, and since 1 < a < n, the
integer 2a − 1 is a proper divisor of 2n − 1 (i.e., strictly greater than 1 and less than n). Hence 2n − 1
is composite.

(ii) Let p and q be distinct primes. Prove that, for any a ∈ Z, apq + a ≡ ap + aq mod pq.

Page 2
Math 453, Section X13 Final Exam Solutions Spring 2011

Solution: This is similar to Problem 59 from HW 4. First observe that, since p and q are distinct
primes, a congruence modulo pq is equivalent to a system of two congruences modulo p and modulo q.
Thus, it suffices to prove that the desired congruence holds modulo each of the primes p and q, i.e., that

(1) aqp + a ≡ aq + ap mod p, aqp + a ≡ aq + ap mod q.

By Fermat’s Little Theorem we have, for any integer a,

(2) a ≡ ap mod p.

Applying this with aq in place of a, we get

(3) aqp = (aq )p ≡ aq mod p.

Substituting (2) and (3) on the left side of (1) gives the first congruence in (1); the second follows in
the same way by interchanging the roles of p and q.

(iii) In the RSA encryption system, a messsage M is encrypted by computing E ≡ M e mod m, where e is
the public encryption exponent and m the public modulus. The encrypted message E is decrypted by
computing E d mod m, where d is the decryption exponent. State precisely how d is defined, and prove
that with this choice of d decryption returns the original message M , i.e., that E d ≡ M mod m.
Solution: The decryption exponent is defined as a (positive) solution to the congruence

de ≡ 1 mod ϕ(m).

With this definition, de = 1 + kϕ(m) for some positive integer k. Since by Euler’s Theorem M ϕ(m) ≡
1 mod m, we get

E d ≡ (M e )d = M ed = M ϕ(m)k+1 = (M ϕ(m) )k M ≡ 1k M ≡ M mod m.

Problem 5
State three primality tests covered in class. In each case, state the name of the test, the precise conditions
that need to be tested, and the conclusion (e.g., “n is prime”, “n is composite”, “test is inconclusive”), for
each of the possible outcomes.
(i) Test 1:
(ii) Test 2:
(iii) Test 3:
Solution: Possible choices are Trial Division, Fermat Test, Lucas Test, Wilson’s Test, and Pepin’s Test.
See the class notes for the precise formulation.

Problem 6

(i) Expand 5/18 into a simple continued fraction, and find all of its convergents Ci .

Page 3
Math 453, Section X13 Final Exam Solutions Spring 2011

Solution: We have
5 1 1 1
=0+ =0+ =0+
18 18 3 1
3+ 3+
5 5 5
3
1 1
=0+ =0+
1 1
3+ 3+
2 1
1+ 1+
3 1
1+
2
= [0, 3, 1, 1, 2]

The convergents are

0
C0 = [0] = ,
1
1 1
C1 = [0, 3] = 0 + = ,
3 3
1 1
C2 = [0, 3, 1] = 0 + = ,
1 4
3+
1
1 2
C3 = [0, 3, 1, 1] = 0 + = ,
1 7
3+
1
1+
1
5
C4 = [0, 3, 1, 1, 2] = .
18

(ii) Find the number α whose simple continued fraction expansion is α = [1, 2, 3, 2, 3, . . . ] = [1, 2, 3].
Solution: We have α = 1 + β1 , where β = [2, 3]. Now,

β = [2, 3, 2, 3] = [2, 3, β]
1 β
=2+ =2+
1 3β + 1
3+
β
7β + 2
= ,
3β + 1
so
3β 2 − 6β − 2 = 0.
Solving this quadratic equation (ignoring the negative root, since β = [2, 3] must be a positive number),

Page 4
Math 453, Section X13 Final Exam Solutions Spring 2011

we get
√ √
6+ 60 3+15
β= = ,
6 3
1 3
α=1+ =1+ √
β 3 + 15
√ √ √
15 + 6 ( 15 + 6)( 15 − 3)
=√ =
15 + 3 15 − 32
√
−1 + 15
= .
2

Problem 7

(i) Evaluate f (1099 ), where f (n) = d|n µ(d)σ(n/d). (With the right approach, this requires only min-
P

imal amount of numerical calculation. Explain any non-obvious steps (e.g., by citing an appropriate
theorems/formulas/properties).
Solution: Using Dirichlet product notation, the given identity for f can be written as f = µ ? σ. We
have σ = 1 ? id, so f = µ ? (1 ? id) = (µ ? 1) ? id = δ ? id = id, by the properties of the Dirichlet product.
Hence f (n) = id(n) = n for all n ∈ N, so f (1099 ) = 1099 .

(ii) Let 1 and id denote the arithmetic functions defined by 1(n) = 1 and id(n) = n for all n ∈ N, and let 1−1
and id−1 denote the Dirichlet product inverses of these functions. Express each of the following functions
as a Dirichlet product of two of these functions, i.e., in the form f ? g, where f, g ∈ {1, 1−1 , id, id−1 }.

(a) (Sum-of-divisors function) σ =

(b) (Euler phi function) ϕ =
(c) (Divisor function) ν =

Solution: (a) σ = 1 ? id; (b) ϕ = id ? 1−1 (since id = ϕ ? 1 by Gauss’ Identity); (c) ν = 1 ? 1.

(iii) Determine, with proof, all positive integers n for which ϕ(3n) = 2ϕ(n).
Solution: Writing n = 3α m, where (3, m) = 1 and α is a nonnegative integer, we have, by the
multiplicativity of ϕ, ϕ(3n) = ϕ(3α+1 m) = ϕ(3α+1 )ϕ(m) and ϕ(n) = ϕ(3α m) = ϕ(3α )ϕ(m). Hence
ϕ(3n) = 2ϕ(n) holds if and only if (∗) ϕ(3α+1 ) = 2ϕ(3α ). The left side of (∗) equals ϕ(3α+1 ) = 2 · 3α
for any α ≥ 0. On the other hand, the right side equals 2 if α = 0, and 2 · 3α−1 if α ≥ 1. Hence (∗)
holds if and only if α = 0. Therefore the positive integers n with ϕ(3n) = 2ϕ(n) are exactly those that
are not divisible by 3.
P
(iv) Let f (n) = d|n ν(d) (where ν(d) is the number-of-divisors function). Find a formula for f (n) in terms
of the standard prime factorization n = pα αr
1 . . . pr (where, as usual, the pi are distinct primes and the
1

αi are positive integers). Use back of page for work if needed.

Solution: Since ν is multiplicative, so is f . At prime powers pα , we can compute f directly from the
definition:
α α
X X X (α + 1)(α + 2)
f (pα ) = ν(d) = ν(pβ ) = (β + 1) = .
2
d|pα β=0 β=0

Page 5
Math 453, Section X13 Final Exam Solutions Spring 2011

By the multiplicativity of f it follows that for n = pα αr

1 . . . pr ,
1

r
Y (αi + 2)(αi + 1)
f (n) = f (pα αr α1 αr
1 . . . pr ) = f (p1 ) . . . f (pr ) =
1

i=1
2

Problem 8

(i) Determine, with explanation, the exact set of integers that are orders modulo 151 of some integer
coprime to 151 (i.e., the set {ord151 a : a ∈ Z, (a, 151) = 1}.)
Solution: The possible orders modulo a prime p are the positive divisors of ϕ(p) = p − 1. Since
p − 1 = 150 = 2 · 3 · 52 , these are the numbers 1, 2, 3, 5, 6, 10, 15, 25, 30, 50, 75, 150 . (Note that there are
ν(150) = ν(21 · 31 · 52 ) = (1 + 1)(1 + 1)(2 + 1) = 12 such divisors.)

(ii) Determine, with explanation, whether there exists an integer n such that n4 + 1 is divisible by 2011.
(Note that 2011 is prime.)
Solution: No such integer exists. Proof (by contradiction): Suppose such an n exists. Then (n2 )2 =
n4 ≡ −1 mod 2011. Therefore (∗) x2 ≡ −1 mod 2011 has a solution (namely x = n2 ), which means that
−1 is a quadratic residue modulo2011.
 However, since 2011 ≡ 3 mod 4, −1 is a quadratic nonresidue
−1
modulo 2011 (by the formula for p ). Thus we have a contradiction.

(iii) Determine, with explanation, whether there exists an integer a in the range 1 < a < 2011 such that
a11 − 1 is divisible by 2011. (Note that 2011 is prime.)
Solution: We claim that no such integer exists. The divisibility condition on a implies that the congru-
ence (∗) ak ≡ 1 mod 2011 holds for k = 11. Thus the order of a modulo 2011 must be a (positive) divisor
of 11, and hence equal to 1 or 11. Since 1 < a < 2011, we have a1 6≡ 1 mod 11, so the order of a cannot
be 1 and therefore must be 11. On the other hand, the order must be a divisor of ϕ(2011) = 2010. Since
11 does not divide 2010 (this can easily be seen by the divisibility test modulo 11), this is impossible.

Problem 9
Extra Credit: Let p be an odd prime, and let r be a primitive root modulo p. Using the known properties
of primitive roots, give a careful proof of Wilson’s Theorem. (You may use other named theorems in your
proof.)
Solution: Wilson’s Theorem states that, when p is prime, then (∗) (p − 1)! ≡ −1 mod p. The key property
of primitive roots needed for the proof is that the numbers r1 , r2 , . . . , rp−1 form a reduced system of residues
modulo p, and thus are congruent modulo p to the numbers 1, 2, . . . , p − 1, in some permutation. Therefore
p−1
Y
(p − 1)! ≡ ri = r1+2+···+(p−1) = r(p−1)p/2 mod p
i=1
p  p

(p−1)/2 r
= r ≡ (by Euler’s Criterion)
p
= (−1)p (since a primitive root is a quadratic nonresidue)
= −1 (since p is odd),

which proves (∗).

Page 6