Number Theory - A

1. What is the remainder, in base 10, when 247 + 3647 + 437 + 127 + 37 + 17 is divided by 6?

(ANS: 3 CB: GBH?, ACH, KB)

2. How many zeros are there at the end of 792! when written in base 10?

(ANS: 196 CB: ACH)

3. Find all integral solutions to xy − y x = 1.

(ANS: (2, 1) and (3, 2) must be given; (x 6= 0, 0) may also be. If both x and y are greater than 3
and not equal, we have |xy − y x | > 2 for analytic reasons. Thus one of x and y are 1 or 2, so you
can check that the two given are the only solutions.
In some versions of the test, the word “positive” was included; therefore the third solution (with
y = 0) was optional. Credit was given to any contestants who answered either version correctly,
that is, who gave either all three answers or just the first two. CB: AL, IAF)

2010 2008
4. Find the largest integer n, where 2009n divides 20082009 + 20102009 .

(ANS: Using the binomial theorem:

2010
A = (2009 − 1)2009 = −1 + 2009 ∗ (20092010 ) + R1
2008
B = (2009 + 1)2009 = 1 + 2009 ∗ (20092008 ) + R2
R1 is divisible by 20092010 and R2 is divisible by 20092010 . A + B ≡ 20092009 (mod 20092010 ), so
n = 2009. CB: AL)

5. How many integers n are there such that 0 ≤ n ≤ 720 and n2 ≡ 1 (mod 720)?

(ANS: 16. 720 = 16 ∗ 9 ∗ 5. There are two solutions mod 5 and mod 9, and there are 4 mod 16,
and so, by the chinese remainder theorem, there are 16 solutions mod 720. But the numbers
1, 2, . . . 720 form a complete reside system mod 720, so there are 16 solutions in all. CB: IAF)

6. f (n) is the sum of all integers less than n and relatively prime to n. Find all integers n such that
there exist integers k and l such that f (nk ) = nl .

1
 Number Theory - A
(ANS: 2, 3, 4, 6. Let φ(n) denote the number of positive integers less than n that are relatively
prime to n. It is clear that 2f (n) = nφ(n), and so we need either φ(n) = 1 or φ(n) = 2, and one
can check that 2, 3, 4, and 6 are the only solutions to this. Moreover, for those values of n, we
have f (n) = n (except 2, where we have f (2) = 1) CB: AL, IAF)

7. In this problem, we consider only polynomials with integer coeffients. Call two polynomials p and
q really close if p(2k + 1) ≡ q(2k + 1) (mod 210 ) for all k ∈ Z+ . Call a polynomial p partial credit if
no polynomial of lesser degree is really close to it. What is the maximum possible degree of partial
credit?

(ANS: 5.
Suppose we have a polynomial p(n) that has the maximum possible degree of partial credit.
Consider q(x) = (x − 1)(x − 3)(x − 5)(x − 7)(x − 9)(x − 11). Observe that for all odd integers x,
q(x) is the product of six consecutive even integers, and thus is divisible by 2 · 4 · 6 · 8 · 10 · 12.
Note that the power of two dividing 2 · 4 · 6 · 8 · 10 · 12 is 210 . Thus we conclude that q(2k + 1) ≡ 0
mod 210 . Now if p(n) has degree six or larger, let its leading term be an` . Then p(n) and
p(n) − an`−6 q(n) are really close, but the latter has lesser degree. Thus the maximum possible
degree is at most five.
Consider p(x) = (x − 1)(x − 3)(x − 5)(x − 7)(x − 9). This is a polynomial of degree five. I claim
that this polynomial is partial credit. Suppose for sake of contradiction that there exists a
polynomial q(x) of degree four or less such that p(2k + 1) ≡ q(2k + 1) mod 210 for all k. We
know q(1) is divisible by 1024. We may add a multiple of 1024 to q(x) so that q(1) = 0 and q is
still really close to p. Thus q(x) = (x − 1)r(x). Now consider q(3). It is divisible by 1024, so r(3)
is divisible by 512. Now adding a multiple of 512 to r(x) does not change the residue class of q(x)
modulo 1024 for x odd. Thus we may assume that r(3) = 0. Hence we get
q(x) = (x − 1)(x − 3)s(x). Similarly, we find s(5) is divisible by 128, and adding multiples of 128
to s(x) does not change the residue class of q(x) modulo 1024 for odd values of x. Hence WLOG
q(x) = (x − 1)(x − 3)(x − 5)h(x). Iterating this argument, we find
q(x) = (x − 1)(x − 3)(x − 5)(x − 7)(x − 9)k(x) for some polynomial k(x). Since q has degree less
than five, we conclude that k is the zero polynomial. Thus we see that p is really close to zero.
But this is clearly false since the largest power of two dividing p(11) = (−2)(−4)(−6)(−8)(−10) is
28 . Thus we have a contradiction. Hence no such polynomial q exists, and thus we conclude that
p is in fact partial credit. CB: ACH,JVP)

xx
8. If f (x) = xx , find the last two digits of f (17) + f (18) + f (19) + f (20).

(ANS: 32. We are interested in finding f (17) + f (18) + f (19) + f (20) modulo 100. By the
Chinese Remainder Theorem it suffices to find the sum modulo 4 and 25.
A ≡ 0 (mod 4) because 18 and 20 are raised to high powers and 19 and 17 are both raised to odd
powers.

2
 Number Theory - A

184 ≡ 1 (mod 25) ⇒ f (18) ≡ 1 (mod 25)

f (20) ≡ 0 (mod 25)

17
To compute f (17) modulo 25, we need to compute X = 1717 modulo 20. X ≡ 1 (mod 4) since
17 ≡ 1 (mod 4). X ≡ 2 (mod 5) and therefore X ≡ 17 (mod 20). Since 3 ∗ 17 ≡ 1 (mod 25),
1717 ≡ 33 (mod 25).

f (17) ≡ 2 (mod 25)

19
To compute f (19) modulo 25, we need to compute Y = 1919 modulo 20.

19 ≡ −1 (mod 20) ⇒ Y ≡ −1 (mod 20)

f (19) ≡ 19−1 ≡ 4 (mod 25)

f (17) + f (18) + f (19) + f (20) ≡ 2 + 1 + 4 + 0 ≡ 7 (mod 25).
f (17) + f (18) + f (19) + f (20) ≡ 0 (mod 4) because 18 and 20 are raised to high powers and 19
and 17 are both raised to odd powers.
f (17) + f (18) + f (19) + f (20) ≡ 0 (mod 4).
The last two digits of f (17) + f (18) + f (19) + f (20) are 32. CB: AL, IAF)

9. What is the largest integer which cannot be expressed as 2008x + 2009y + 2010z for some positive
integers x, y, and z?

(ANS: 2016031. The largest number that cannot be a positive linear combanation of 1004 and
1005 is 1004 ∗ 1005 − 1004 − 1005 = 1007011, so the smallest even number that cannot be written
as the sum of 2008 and 2010 is 2007022. Then the largest number that we can’t get is 2009031, as
every larger number you can make it even by subtracting 2009 if necessary and write that as the
sum of multiples of 2008 and 2010. CB: AL12, IAF)

10. Find the smallest positive integer n such that 32n = 167x + 2 for some integer x.

(ANS: 50. 3250 = 13500 = 132+3∗166 = 132 = 2, as 167 is prime. Since 2 is a perfect square, and
nothing else in the unit group, the answer is less than 83, and so it is 50.
Alternately, 32 = 25 . 283 ≡ 1 (mod 167). So we need the smallest n such that 5n ≡ 1 (mod 83).
That gives n = 50. CB: EK)

3
 Number Theory - A
11. Find all sets of three primes p, q, and r such that p + q = r and (r − p)(q − p) − 27p is a perfect
square.

(ANS: 2, 29, 31. It is easy to see that q > p and that one of p and q is two, so p = 2. We then
have q(q − 2) − 54 = x2 , so (q − 1)2 − x2 = 55. Thus, since the prime factorization of 55 is 5 ∗ 11,
we have either q − 1 = 8 and x = 3, or q − 1 = 28 and x = 27. Only q = 29 has the solution be
only primes. CB: AL, IAF)

12. Find the number of positive integer solutions of (x2 + 2)(y 2 + 3)(z 2 + 4) = 60xyz.

(ANS: 8. We have that xyz < 60. We also have that, if a solution uses z = 1, then you can
replace that with z = 4, and similarly, you can replace y = 1 with y = 3, and x = 1 with x = 2.
Moreover, you have that z ≡ ±1(5), as there is no way to force 5 to divide x2 + 2 or y 2 + 3. Since
any solution of the equation can be forced to have x, y ≥ 2, we have to only check
z = 1, 4, 6, 9, 11, and14. The only possible solution for z = 14 is x = y = 2with the already stated
equivalences, and one can check that that fails. If z = 11, we need 2 powers of 5 on the right side
of the equation, which forces xyz to be too big. If z = 9, we have that 17|xy and so xyz is too
big. If z = 6, we have that (x2 + 2)(y 2 + 3) = 9xy. Thus xy < 9 and since ν3 (y) ≥ ν3 (y 2 + 3), we
have x = 4, 5, and, since we can force y to be at least 2, we have x = 4 and y = 2, which doesn’t
work. Thus, we have z = 4 (or 1). The work here is similar and similarly tedious, and the only
solution are the 8 that are equivalent to x = y = z = 1. CB: AL)

13. What is the smallest number n such that you can choose n distinct odd integers a1 , a2 , . . . an , none
of them 1, with a11 + a12 + . . . + a1n = 1?

(ANS: 9. You can check that the answer is not less than 8, as 13 + 15 + 17 + 91 + 11 1
+ 13 1 1
+ 15 < 1.
You also can check that it is not 8, as the sum is, once you normalize denominators, the sum of
odd numbers over an odd number. Thus, we need to show that 9 works. To see that 9 works, you
can choose 3, 5, 7, 9, 11, 15, 35, 45, 231 as the ai . CB: IAF)