Questions based on Wilson’s theorem for training for RMO

1(a) Find the remainder when 15! is divided by 17.
1(b) Find the remainder when 2(26!) is divided by 29.

2: Determine whether 17 is a prime by deciding if 16! \equiv -1 {\pmod 17}

3: Arrange the integers 2,3,4, …, 21 in pairs a and b that satisfy ab \equiv 1 {\pmod 23}.

4: Show that 18! \equiv -1 {\pmod 437}.

5a: Prove that an integer n>1 is prime if and only if (n-2)! \equiv 1 {\pmod n}.
5b: If n is a composite integer, show that (n-1)! \equiv 0 {\pmod n}, except when n=4.

6: Given a prime number p, establish the congruence (p-1)! \equiv {p-1} {\pmod {1+2+3+\ldots + (p-1)}}

7: If p is prime, prove that for any integer a, p|a^{p}+(p-1)|a and p|(p-1)!a^{p}+a

8: Find two odd primes p \leq 13 for which the congruence (p-1)! \equiv -1 {\pmod p^{2}} holds.

9: Using Wilson’s theorem, prove that for any odd prime p:
1^{2}.3^{2}.5^{2}.\ldots (p-2)^{2} \equiv (-1)^{(p+1)/2} {\pmod p}

10a: For a prime p of the form 4k+3, prove that either

(\frac{p-1}{2})! \equiv 1 {\pmod p} or (\frac{p-1}{2})! \equiv -1 {\pmod p}

10b: Use the part (a) to show that if 4k+3 is prime, then the product of all the even integers less than p is congruent modulo p to either 1 or -1.

More later,
Nalin Pithwa.

Wilson’s theorem and related problems in Elementary Number Theory for RMO

I) Prove Wilson’s Theorem:

If p is a prime, then (p-1)! \equiv -1 {\pmod p}.


The cases for primes 2 and 3 are clearly true.

Assume p>3

Suppose that a is any one of the p-1 positive integers 1,2,3, \ldots {p-1} and consider the linear congruence
ax \equiv 1 {\pmod p}. Then, gcd(a,p)=1.

Now, apply the following theorem: the linear congruence ax \equiv b {\pmod n} has a solution if and only if d|b, where d = gcd(a,b). If d|b, then it has d mutually incongruent solutions modulo n.

So, by the above theorem, the congruence here admits a unique solution modulo p; hence, there is a unique integer a^{'}, with 1 \leq a^{'} \leq p-1, satisfying aa^{'} \equiv 1 {\pmod p}.

Because p is prime, a = a^{'} if and only if a=1 or a=p-1. Indeed, the congruence a^{2} \equiv 1 {\pmod p} is equivalent to (a-1)(a+1) \equiv 0 {\pmod p}. Therefore, either a-1 \equiv 0 {\pmod p}, in which case a=1, or a+1 \equiv 0 {\pmod p}, in which case a=p-1.

If we omit the numbers 1 and p-1, the effect is to group the remaining integers 2,3, \ldots (p-2) into pairs a and a^{'}, where a \neq a^{'}, such that the product aa^{'} \equiv 1 {\pmod p}. When these (p-3)/2 congruences are multiplied together and the factors rearranged, we get

2.3. \ldots (p-2) \equiv 1 {\pmod p}

or rather

(p-2)! \equiv 1 {\pmod p}

Now multiply by p-1 to obtain the congruence

(p-1)! \equiv p-1 \equiv -1 {\pmod p}, which was desired to be proved.

An example to clarify the proof of Wilson’s theorem:

Specifically, let us take prime p=13. It is possible to divide the integers 2,3,4, \ldots, 11 into (p-3)/2=5 pairs, each product of which is congruent to 1 modulo 13. Let us write out these congruences explicitly as shown below:

2.7 \equiv 1 {\pmod {13}}
3.9 \equiv 1 {\pmod {13}}
4.10 \equiv 1 {\pmod {13}}
5.8 \equiv 1 {\pmod {13}}
6.11 \equiv 1 {\pmod {13}}

Multpilying these congruences gives the result 11! = (2.7)(3.9)(4.10)(5.8)(6.11) \equiv 1 {\pmod {13}}

and as 12! \equiv 12 \equiv -1 {\pmod {13}}

Thus, (p-1)! \equiv -1 {\pmod p} with prime p=13.


The converse to Wilson’s theorem is also true. If (n-1)! \equiv -1 {\pmod n}, then n must be prime. For, if n is not a prime, then n has a divisor d with 1 1 is prime if and only if (n-1)! \equiv -1 {\pmod n}. Unfortunately, this test is of more theoretical than practical interest because as n increases, (n-1)! rapidly becomes unmanageable in size.

Let us illustrate an application of Wilson’s theorem to the study of quadratic congruences{ What we mean by quadratic congruence is a congruence of the form ax^{2}+bx+c \equiv 0 {\pmod n}, with a \not\equiv 0 {\pmod n} }

Theorem: The quadratic congruence x^{2}+1 \equiv 0 {\pmod p}, where p is an odd prime, has a solution if and only if p \equiv 1 {\mod 4}.


Let a be any solution of x^{2}+1 \equiv 0 {\pmod p} so that a^{2} \equiv -1 {\pmod p}. Because p \not |a, the outcome of applying Fermat’s Little Theorem is

1 \equiv a^{p-1} \equiv (a^{2})^{(p-1)/2} \equiv (-1)^{(p-1)/2} {\pmod p}

The possibility that p=4k+3 for some k does not arise. If it did, we would have

(-1)^{(p-1)/2} = (-1)^{2k+1} = -1

Hence, 1 \equiv -1 {\pmod p}. The net result of this is that p|2, which is clearly false. Therefore, p must be of the form 4k+1.

Now, for the opposite direction. In the product

(p-1)! = 1.2 \ldots \frac{p-1}{2} \frac{p+1}{2} \ldots (p-2)(p-1)

we have the congruences

p-1 \equiv -1 {\pmod p}
p-2 \equiv -2 {\pmod p}
p-3 \equiv -3 {\pmod p}
\frac{p+1}{2} \equiv - \frac{p-1}{2} {\pmod p}

Rearranging the factors produces
(p-1)! \equiv 1.(-1).2.(-2) \ldots \frac{p-1}{2}. (-\frac{p-1}{2}) {\pmod p} \equiv (-1)^{(p-1)/2}(.2. \ldots \frac{p-1}{2})^{2}{\pmod p}

because there are (p-1)/2 minus signs involved. It is at this point that Wilson’s theorem can be brought to bear; for, (p-1)! \equiv -1 {\pmod p}, hence,

-1 \equiv (-1)^{(p-1)/2}((\frac{p-1}{2})!)^{2} {\pmod p}

If we assume that p is of the form 4k+1, then (-1)^{(p-1)/2} =1, leaving us with the congruence

-1 \equiv (-\frac{p-1}{2})^{2}{\pmod p}.

The conclusion is that the integer (\frac{p-1}{2})! satisfies the quadratic congruence x^{2}+1 \equiv 0 {\pmod p}.

Let us take a look at an actual example, say, the case p=13, which is a prime of the form 4k+1. Here, we have \frac{p-1}{2}=6, and it is easy to see that 6! = 720 \equiv 5 {\pmod {13}} and 5^{2}+1 = 26 \equiv 0 {\pmod {13}}.

Thus, the assertion that ((p-1)!)^{2}+1 \equiv 0 {\pmod p} is correct for p=13.

Wilson’s theorem implies that there exists an infinitude of composite numbers of the form n!+1. On the other hand, it is an open question whether n!+1 is prime for infinitely many values of n. Refer, for example:

More later! Happy churnings of number theory!
Nalin Pithwa

Eight digit bank identification number and other problems of elementary number theory

Question 1:

Consider the eight-digit bank identification number a_{1}a_{2}\ldots a_{8}, which is followed by a ninth check digit a_{9} chosen to satisfy the congruence

a_{9} \equiv 7a_{1} + 3a_{2} + 9a_{3} + 7a_{4} + 3a_{5} + 9a_{6} + 7a_{7} + 3a_{8} {\pmod {10}}

(a) Obtain the check digits that should be appended to the two numbers 55382006 and 81372439.

(b) The bank identification number 237a_{4}18538 has an illegitimate fourth digit. Determine the value of the obscured digit.

Question 2:

(a) Find an integer having the remainders 1,2,5,5 when divided by 2, 3, 6, 12 respectively (Yih-hing, died 717)

(b) Find an integer having the remainders 2,3,4,5 when divided by 3,4,5,6 respectively (Bhaskara, born 1114)

(c) Find an integer having remainders 3,11,15 when divided by 10, 13, 17, respectively (Regiomontanus, 1436-1476)

Question 3:

Question 3:

Let t_{n} denote the nth triangular number. For which values of n does t_{n} divide t_{1}^{2} + t_{2}^{2} + \ldots + t_{n}^{2}

Hint: Because t_{1}^{2}+t_{2}^{2}+ \ldots + t_{n}^{2} = t_{n}(3n^{3}+12n^{2}+13n+2)/30, it suffices to determine those n satisfying 3n^{3}+12n^{2}+13n+2 \equiv 0 {\pmod {2.3.5}}

Question 4:

Find the solutions of the system of congruences:

3x + 4y \equiv 5 {\pmod {13}}
2x + 5y \equiv 7 {\pmod {13}}

Question 5:

Obtain the two incongruent solutions modulo 210 of the system

2x \equiv 3 {\pmod 5}
4x \equiv 2 {\pmod 6}
3x \equiv 2 {\pmod 7}

Question 6:

Use Fermat’s Little Theorem to verify that 17 divides 11^{104}+1

Question 7:

(a) If gcd(a,35)=1, show that a^{12} \equiv {\pmod {35}}. Hint: From Fermat’s Little Theorem, a^{6} \equiv 1 {\pmod 7} and a^{4} \equiv 1 {\pmod 5}

(b) If gcd(a,42) =1, show that 168=3.7.8 divides a^{6}-1
(c) If gcd(a,133)=gcd(b,133)=1, show that 133| a^{18} - b^{18}

Question 8:

Show that 561|2^{561}-1 and 561|3^{561}-3. Do there exist infinitely many composite numbers n with the property that n|2^{n}-2 and n|3^{n}-3?

Question 9:

Prove that any integer of the form n = (6k+1)(12k+1)(18k+1) is an absolute pseudoprime if all three factors are prime; hence, 1729=7.13.19 is an absolute pseudoprime.

Question 10:

Prove that the quadratic congruence x^{2}+1 \equiv 0 {\pmod p}, where p is an odd prime, has a solution if and only if p \equiv {pmod 4}.

Note: By quadratic congruence is meant a congruence of the form ax^{2}+bx+c \equiv 0 {\pmod n} with a \equiv 0 {\pmod n}. This is the content of the above proof.

More later,
Nalin Pithwa.

Elementary Number Theory, ISBN numbers and mathematics olympiads

Question 1:

The International Standard Book Number (ISBN) used in many libraries consists of nine digits a_{1} a_{2}\ldots a_{9} followed by a tenth check digit a_{10} (somewhat like Hamming codes), which satisfies

a_{10} = \sum_{k=1}^{9}k a_{k} \pmod {11}

Determine whether each of the ISBN’s below is correct.
(a) 0-07-232569-0 (USA)
(b) 91-7643-497-5 (Sweden)
(c) 1-56947-303-10 (UK)

Question 2:

When printing the ISBN a_{1}a_{2}\ldots a_{9}, two unequal digits were transposed. Show that the check digits detected this error.

Remark: Such codes are called error correcting codes and are fundamental to wireless communications including cell phone technologies.

More later,
Nalin Pithwa.

A good way to start mathematical studies …

I would strongly suggest to read the book “Men of Mathematics” by E. T. Bell.

It helps if you start at a young age. It doesn’t matter if you start later because time is relative!! 🙂

Well, I would recommend you start tinkering with mathematics by playing with nuggets of number theory, and later delving into number theory. An accessible way for anyone is “A Friendly Introduction to Number Theory” by Joseph H. Silverman. It includes some programming exercises also, which is sheer fun.

One of the other ways I motivate myself is to find out biographical or autobiographical sketches of mathematicians, including number theorists, of course. In this, the internet is an extremely useful information tool for anyone willing to learn…

Below is a list of some famous number theorists, and then there is a list of perhaps, not so famous number theorists — go ahead, use the internet and find out more about number theory, history of number theory, the tools and techniques of number theory, the personalities of number theorists, etc. Become a self-learner, self-propeller…if you develop a sharp focus, you can perhaps even learn from MIT OpenCourseWare, Department of Mathematics.

Famous Number Theorists (just my opinion);

1) Pythagoras
2) Euclid
3) Diophantus
4) Eratosthenes
5) P. L. Tchebycheff (also written as Chebychev or Chebyshev).
6) Leonhard Euler
7) Christian Goldbach
8) Lejeune Dirichlet
9) Pierre de Fermat
10) Carl Friedrich Gauss
11) R. D. Carmichael
12) Edward Waring
13) John Wilson
14) Joseph Louis Lagrange
15) Legendre
16) J. J. Sylvester
11) Leonoardo of Pisa aka Fibonacci.
15) Srinivasa Ramanujan
16) Godfrey H. Hardy
17) Leonard E. Dickson
18) Paul Erdos
19) Sir Andrew Wiles
20) George Polya
21) Sophie Germain
24) Niels Henrik Abel
25) Richard Dedekind
26) David Hilbert
27) Carl Jacobi
28) Leopold Kronecker
29) Marin Mersenne
30) Hermann Minkowski
31) Bernhard Riemann

Perhaps, not-so-famous number theorists (just my opinion):
1) Joseph Bertrand
2) Regiomontanus
3) K. Bogart
4) Richard Brualdi
5) V. Chvatal
6) J. Conway
7) R. P. Dilworth
8) Martin Gardner
9) R. Graham
10) M. Hall
11) Krishnaswami Alladi
12) F. Harary
13) P. Hilton
14) A. J. Hoffman
15) V. Klee
16) D. Kleiman
17) Donald Knuth
18) E. Lawler
19) A. Ralston
20) F. Roberts
21) Gian Carlo-Rota
22) Bruce Berndt
23) Richard Stanley
24) Alan Tucker
25) Enrico Bombieri

Happy discoveries lie on this journey…
-Nalin Pithwa.

Any integer can be written as the sum of the cubes of 5 integers, not necessarily distinct

Question: Prove that any integer can be written as the sum of the cubes of five integers, not necessarily.


We use the identity 6k = (k+1)^{3} + (k-1)^{3}- k^{3} - k^{3} for k=\frac{n^{3}-n}{6}=\frac{n(n-1)(n+1)}{6}, which is an integer for all n. We obtain

n^{3}-n = (\frac{n^{3}-n}{6}+1)^{3} + (\frac{n^{3}-n}{6}-1)^{3} - (\frac{n^{3}-n}{6})^{3} - (\frac{n^{3}-n}{6}).

Hence, n is equal to the sum

(-n)^{3} + (\frac{n^{3}-n}{6})^{3} + (\frac{n^{3}-n}{6})^{3} + (\frac{n-n^{3}}{6}-1)^{3}+ (\frac{n-n^{3}}{6}+1)^{3}.

More later,
Nalin Pithwa.