This action might not be possible to undo. Are you sure you want to continue?

# MATH 115A SOLUTION SET I JANUARY 13, 2005 1 (i) Suppose that n > 1 is a composite integer, with n = rs, say.

Show that 2n − 1 is divisible by 2r − 1. (This shows that 2n − 1 is prime only if n is prime. Primes of the form 2n − 1 are called Mersenne primes.) (ii) Show that if 2k + 1 is prime, then k must be a power of 2. (This explains why Fermat n only had to consider numbers of the form fn = 22 + 1 (using the notation described in class) when he was hunting for primes, rather than more general numbers of the form 2k + 1 for an arbitrary positive integer k.) Solution: Suppose that n = rs. Then we have: 2rs − 1 = (2r )s − 1 = [2r − 1][(2r )s−1 + ... + 2r + 1]. Hence Mr divides Mn , as claimed. To see why Fermat could restrict himself to the numbers fn instead of considering the more general class of numbers 2k + 1, it suﬃces to show that if 2k + 1 is a prime, then k must be a power of two. Suppose that p is an odd prime factor of k, so k = mp, say. Then 2k + 1 = (2m + 1)(1 − 2m + 22m − ... + 2(p−1)m ). Hence, if k is divisible by an odd prime, then 2k + 1 is composite, and this establishes the result we seek. (2) The Fibonacci sequence (Fn )n≥1 is deﬁned recursively by the equations F1 = 1, F2 = 1, and Fn+1 = Fn + Fn−1 for n ≥ 2. √ Let α be any real number larger than β = 1 (1 + 5). Prove by induction or otherwise 2 that, for n ≥ 1, Fn < αn . [Hint: You may ﬁnd it helpful to note that β is a solution of the equation x2 = x + 1.] Solution: We prove this result via induction. √ First observe that the two roots of the quadratic equation x2 = x + 1 are β = (1 + 5)/2 √ and β1 = (1 − 5)/2, and we have that β > β1 . Hence it follows that if α is any real number strictly greater than β, then α2 > α + 1. Now suppose that α is a ﬁxed real number strictly greater than β. Let P (n) denote the statement ‘Fn < αn ’. Since F1 = F2 = 1, it follows that P (1) and P (2) are true.

1

Since 1 = 1(2)/2. Then it follows from the Pigeonhole Principle that in any collection of n+1 individual gloves. Let P (n) denote the statement ‘ n m = n(n + 1)/2’. This is plainly the smallest number that will suﬃce. Hence if we remove n + 1 gloves from the drawer. it follows that 1 P (1) is true. then so is P (k + 1). (3) Prove that. Hence if P (k) is true. and so it follows by induction that P (n) is true for all positive integers n. and these two gloves yield a complete pair. 2 Solution: We prove this result via induction. n individual right-hand gloves that do not form a complete pair. since it is certainly possible to have a collection of e. [Hint: 211 − 1 = 23 × 89. at least two gloves must be labelled with the same integer.g.2 MATH 115A SOLUTION SET I JANUARY 13. Now let k be an integer for which P (k) is true. = (4) Suppose that n pairs of gloves of diﬀerent sizes are mixed together in a drawer. 2 Hence if P (k) is true. Then k+1 k m= 1 1 m + (k + 1) k(k + 1) + (k + 1) (by our inductive hypothesis) 2 (k + 1)(k + 2) = . How many individual gloves must you take out if you are to be sure of having at least one complete pair. for n ≥ 1. then we can be sure of having at least one complete pair. 2005 Let k be a positive integer such that P (n) is true for all n ≤ k. and so it follows by induction that P (n) is true for all positive integers n. (5) Prove that the integer n = 210 (211 − 1) is not a perfect number by showing that σ(n) = 2n.] Solution: . Imagine that each individual glove is labelled with an integer between 1 and n to indicate the pair to which it belongs. n (by deﬁnition) (by our inductive hypothesis) 1 1 m = n(n + 1). then so is P (k + 1). Then Fk+1 = Fk + Fk−1 < αk + αk−1 = αk−1 (α + 1) < αk+1 (since α2 > α + 1). (Of course you must justify your answer!) Solution: There are n pairs of gloves.

MATH 115A SOLUTION SET I JANUARY 13. i. p−1 and so it follows that pa cannot be perfect. (6) Verify each of the following statements: (a) No power of a prime can be a perfect number.. (b) The product of two odd primes is never a perfect number.. Solution: If n is perfect. and that a is a positive integer. σ(n) = (1 + 2 + 22 + . since p and q are odd primes. then σ(n) = 2n. (ii) Show that no proper divisor of a perfect number can be perfect. d = 2n. prove that d|n (1/d) = 2. 2005 3 We have n = 210 (211 − 1) = 210 × 23 × 89. + 210 )(1 + 23)(1 + 89) = (211 − 1) × 24 × 90. while 2n = 2 × 210 (211 − 1) = 211 (211 − 1).. n d|n . and so it follows that pq is perfect if and only if 1 + p + q = pq.e. Solution: (a) Suppose that p is a prime.. we have that (p − 1)(q − 1) > 2. This implies that pq > 1 + p + q. (7) (i) If n is a perfect number. Therefore σ(n) = 2n. Then pq is perfect if and only if σ(pq) = 2pq. Then pa is perfect if and only if σ(pa ) = pa . Hence. (b) Suppose that p and q are two distinct odd primes... However. and so n is not perfect. It therefore follows that pq cannot be perfect. + pa−1 = pa − 1 < pa . d|n Multiplying both sides of this equation by 1/n yields d = 2. We now observe that. Now σ(pq) = (1 + p)(1 + q) = 1 + p + q + pq. This in turn holds if and only if 1 + p + p2 + . + pa−1 = pa . 1 + p + p2 + .

d the result follows. + qj j ) b = σ(mn) (a) Since 3 n. q1 . and 3 n. Solution: Let us ﬁrst make the following preliminary observation. qs are distict primes... . Then r s b b n = q11 . using (i) above. and let m be any proper divisor of n. we have that σ(3n) = 4k × 3n.. It therefore follows that m cannot be perfect.. then 5n is 6-perfect. . we have that 1 1 < d d d|m d|n = 2. we have σ(3n) = σ(3)σ(n) = (1 + 3) × 3n = 4 × 3n. since 3 n. 2005 Since d|n d = n d|n 1 . (8) If σ(n) = kn. (c) Since 3n is 4k-perfect. (ii) Suppose that n is a perfect number. we have σ(3n) = (1 + 3)σ(n) = 4σ(n). 1 r where p1 . (b) Since 5 n. and that their prime factorisations are given by m = pa1 . then the positive integer n is called a k-perfect number. Hence 3n is 4-perfect. Establish the following assertions concerning k-perfect numbers: (a) If n is a 3-perfect number..4 MATH 115A SOLUTION SET I JANUARY 13. (c) If 3n is a 4k-perfect number. . σ(m)σ(n) = i=1 (1 + pi + ... and 5 n.... Then. pr . (b) If n is a 5-perfect number.qss . + p ai ) i j=1 (1 + qj + .par .. where k ≥ 3. we have σ(5n) = σ(5)σ(n) = (1 + 5) × 5n = 6 × 5n. then n is 3k-perfect. Suppose that m and n are relatively prime positive integers.. then 3n is 4-perfect. Hence 5n is 6-perfect. On the other hand. and 3 n.

. i. 2005 5 Hence it follows that 4σ(n) = 4k × 3n. This implies that σ(n) = 3k × n.MATH 115A SOLUTION SET I JANUARY 13.e that n is 3k-perfect.