Solutions to Exercises
In this chapter we present solutions to some exercises along with hints for helping to solve others.
Solutions to Exercises in Chapter I
Exercise 1.8. We prove first the validity of the associative law of addition, that is, the equality
n + (m + p) = (n + m) + p (1) for all n, m, p N. We do this by induction on p. For p = 0, the assertion is ∈ clear. Suppose, then, that assertion (1) holds for an arbitrary but fixed p N ∈ and for all n, m N. Then from the definition of addition and the induction hypothesis, we have∈ the equality
(n + m) + p∗ = ((n + m) + p)∗ = (n + (m + p))∗ = n + (m + p)∗ = n + (m + p∗), as desired. This proves the associativity of addition. The first distributive law, that is, the equality
(n + m) p = n p + m p (2) · · · for all n, m, p N, can be proved using the associativity and commutativity ∈ of addition along with induction on p. For p = 0, the assertion is clear. Sup- pose now that (2) holds for an arbitrary but fixed p N and for all n, m N. Then we have ∈ ∈
(n + m) p∗ = (n + m) p + (n + m) = (n p + m p) + (n + m) · · · · = (n p + n) + (m p + m) = n p∗ + m p∗, · · · · as desired. This proves the asserted distributivity. Using the first distributive law, one can then prove the commutativity of multiplication by induction. This then also yields the validity of the second distributive law. Finally, one can prove the associativity of multiplication by induction.
Exercise 1.10. Let m, n N. If m = 0 or n = 0, then one immediately has ∈ the equality m n = 0 by Definition 1.5 (2) and the commutativity of mul- · © Springer International Publishing AG 2017 J. Kramer and A.-M. von Pippich, From Natural Numbers to Quaternions, Springer Undergraduate Mathematics Series, https://doi.org/10.1007/978-3-319-69429-0 248 Solutions to Exercises tiplication. To prove the converse, we assume m = 0 and n = 0 and prove 6 6 the inequality m n = 0. Since m = 0 and n = 0, there exist a,b N with · 6 6 6 ∈ m = a∗ = a + 1 and n = b∗ = b + 1. Therefore, m n = m b∗ = (m b) + m = (m b) + (a + 1) = (m b + a) + 1 = (m b + a)∗, · · · · · · that is, the natural number m n is the successor of the natural number · m b + a. By the third Peano axiom, we must then have m n = 0. · · 6 Exercise 1.14. The power law from Lemma 1.13 can by proved by induction.
Exercise 1.17. The proof of properties (i), (ii), and (iii) of Remark 1.16 are left to the reader.
Exercise 1.20. Properties (i) and (ii) of Remark 1.19 can be proved by induc- tion.
Exercise 1.23. We leave it to the reader to come up with suitable examples.
Exercise 1.25. Let m, n N with m n. We first prove the existence of a ∈ ≥ natural number x N with n + x = m. If m = n, then for x = 0, we have the ∈ equality n + x = n + 0 = m. If m > n, then there exists a N, a > 0, such ∈ that m = n∗···∗ (a times). Then for x = 0∗···∗ (a times), we have the equality n + x = n + 0 = n = m. To prove uniqueness, let y N be another ∗···∗ ∗···∗ ∈ natural number with n + y = m. If x < y, then by Remark 1.19 (i) and the commutativity of addition, we have the inequality
m = n + x = x + n < y + n = n + y = m, that is, we conclude that m < m, a contradiction. If x > y, then there fol- lows analogously the contradiction m > m. Therefore, we must have x = y, proving the asserted uniqueness.
Exercise 2.5. (a) By assumption we have 3 (a a + 1), that is, there exists | 1 ··· k n N with a a + 1 = 3 n. If now 3 a for some j 1,. . .,k , then ∈ 1 ··· k · | j ∈ { } by Lemma 2.4 (ix), we have 3 (a a ), that is, there exists m N with | 1 ··· k ∈ a a = 3 m. We therefore have the equality 1 ··· k · 1 = 3 n a a = 3 n 3 m = 3 (n m), · − 1 ··· k · − · · − a contradiction. Thus none of the numbers a1,..., ak is divisible by 3. (b) We assume that none of the numbers a1 + 1,. . ., ak + 1 is divisible by 3. One first shows that then one must have a + 1 = 3 n + r for certain n N j · j j j ∈ and r 1,2 (j = 1,. . .,k), from which follows a = 3 n + (r 1), which j ∈ { } j · j j − for j = 1,. . .,k implies the equality rj = 2, since by part (a), no aj is divisible by 3. By multiplying out, one shows that the number Solutions to Exercises in Chapter I 249
k a1 ak 1 = ∏(3 nj + 1) 1 ··· − j=1 · − is divisible by 3. But this contradicts the assumption that a a + 1 is di- 1 ··· k visible by 3. Therefore, at least one of the numbers a1 + 1,. . ., ak + 1 must be divisible by 3.
Exercise 2.12. To make the proceedings clear, we calculate first, with a1 = 2 and a = (a 1) a + 1 for n N, n 1, the numbers a = 3, a = 7, n+1 n − · n ∈ ≥ 2 3 a = 43, a = 1807, a = (1807 1) 1806 + 1 = 3263443. With = p P 4 5 6 − · Mn { ∈ | p a , we obtain the first six sets, | n} = 2 , = 3 , = 7 , = 43 , M1 { } M2 { } M3 { } M4 { } = 13, 139 , = 3263443 . M5 { } M6 { } We claim that we have the equality a = 5 b + r with b N and r 2,3 n · n n n ∈ n ∈ { } (n N, n 1). This can be proved by induction on n. If n = 1, the assertion is ∈ ≥ clear. We now assume that the assertion holds for arbitrary but fixed n N, ∈ n 1. Then we have ≥ a = (a 1) a + 1 = (5 b + r 1) (5 b + r ) + 1 n+1 n − · n · n n − · · n n = 5 (5b2 + 2b r b ) + r2 r + 1. · n n n − n n − n If r = 2, then r2 r + 1 = 3; if r = 3, then r2 r + 1 = 5 + 2. Therefore, in n n − n n n − n both cases, a is of the form 5 b + r with b N and r 2,3 , n+1 · n+1 n+1 n+1 ∈ n+1 ∈ { } as desired. We have thus shown that 5 - a , that is, 5 / , for all n N, n ∈ Mn ∈ n 1. ≥ Exercise 2.13. Let us assume, contrary to the assertion, that there are only N finitely many prime numbers p1,..., pn in 2 + 3 . We then consider the natural number · a := 3 p p 1. · 1 ··· n − We have that a > 1, and by Lemma 2.9, a has a prime divisor p. Since 3 - a, it follows that p = 3. We now show that p 2 + 3 N; that is, we must show 6 ∈ · that 3 (p + 1). If p = a, we are done. If p < a, then there exists q N, q > 1, | ∈ with a = p q. Since 3 (p q + 1), it follows by Exercise 2.5 (b) that 3 (p + 1) · | · | or 3 (q + 1). In the first case, we are done. In the second case, we repeat the | process for a prime divisor of q. Finally, after finitely many steps, we obtain a prime divisor p of a with p 2 + 3 N. We now proceed as in Euclid’s proof, for on the assumption that∈ there· are only finitely many prime num- bers in the set 2 + 3 N, we must have p p ,..., p . In particular, we have · ∈ { 1 n} p (p p ). However, since we have the divisibility relation p a, we must | 1 ··· n | have p 1 from the divisibility rules, which is a contradiction. | Exercise 2.15. We prove the contrapositive of the asserted implication. 250 Solutions to Exercises
(i) Suppose that n is not a prime. Then there exist natural numbers a,b N ∈ with n = a b and 1 < a,b < n. We thus obtain · n a b a a (b 1) a (b 2) a 2 1 = 2 · 1 = (2 1) (2 · − + 2 · − + + 2 + 1). − − − · ··· Since 1 < 2a 1 < 2n 1, it follows that 2a 1 is a nontrivial divisor of 2n 1, which− proves that− 2n 1 is not prime. − − − (ii) Let n N, n > 0, with n not a power of 2. Then n > 2, and there exist ∈ natural numbers a,b N, b odd, with n = a b and 1 a < n, 1 < b n. ∈ · ≤ ≤ Since b is odd, we obtain
n a b a a (b 1) a 2 + 1 = 2 · + 1 = (2 + 1) (2 · − 2 + 1). · ∓ · · · − Since 1 < 2a + 1 < 2n + 1, it follows that 2a + 1 is a nontrivial divisor of 2n + 1, which proves that 2n + 1 is not prime.
Exercise 2.18. Let a ,..., a and b ,..., b denote the sets of all divisors { 1 n} { 1 m} of a and all divisors of b. Since a and b are relatively prime, the set of all divisors of a b is equal to a b j = 1,. . ., n; k = 1,. . ., m . It follows that · { j · k | } n m S(a) S(b) = (a1 + + an) (b1 + + bm) = ∑ ∑ aj bk = S(a b). · ··· · ··· j=1 k=1 · ·
This completes the proof of the assertion.
Exercise 2.19. The assertion can be proved by induction on m.
Exercise 2.20. (a) We have the equalities
S(220) 220 = 1 + 2 + 4 + 5 + 10 + 11 + 20 + 22 + 44 + 55 + 110 = 284, − S(284) 284 = 1 + 2 + 3 + 71 + 142 = 220. − Therefore, we have S(220) = 220 + 284 = S(284), which proves that the numbers 220 and 284 are amicable. (b) We must show that S(a) = a + b = S(b). Since x,y,z are distinct odd primes, it follows that
S(a) = S(2n x y) = S(2n) S(x) S(y) = (2n+1 1)(x + 1)(y + 1), · · · · − S(b) = S(2n) S(z) = (2n+1 1)(z + 1), · − where we have used the well-known equality S(2n) = 2n+1 1. A direct − calculation shows that x y = 9 22n 1 9 2n 1 + 1, and therefore, x y + · · − − · − · x + y = z. This implies S(a) = (2n+1 1)(z + 1) = S(b). Finally, we calculate − Solutions to Exercises in Chapter I 251
n n 2n n 1 2n 1 n+1 a + b = 2 (x y + z) = 2 (9 2 9 2 − ) = 2 − 9 (2 1) · · · · − · · · − = (z + 1) (2n+1 1) = S(a) = S(b), · − which proves that the numbers a and b are amicable.
Exercise 3.2. We obtain the prime factorizations 720 = 24 32 5, 9797 = 97 101 and 360360 = (23 32 5)360 = 21080 3720 5360. Finally,· on using· the third· binomial formula four· times,· we obtain· ·
232 1 = (22 1) 22 + 1 24 + 1 28 + 1 216 + 1 − − · · · · = 3 5 17 257 65537. · · · · Exercise 3.7. Let a = 232 1 and b = 255 with prime factorizations (see Ex- ercise 3.2) −
a = ∏ pap = 3 5 17 257 65537, b = ∏ pbp = 3 5 17; p P · · · · p P · · ∈ ∈ here a = 1, a = 1, a = 1, a = 1, a = 1, a = 0 for all p P 3 5 17 257 65537 p ∈ \ 3,5,17,257,65537 and b = 1, b = 1, b = 1, b = 0 for all p P 3,5,17 . { } 3 5 17 p ∈ \{ } Therefore, b a for all p P, which by the criterion of Lemma 3.5, proves p ≤ p ∈ that b a. | Exercise 4.6. With the help of Theorem 4.3, we obtain (3600, 3240) = 360, (360360, 540180) = ((23 32 5)360,(22 33 5)180) = 2360 3540 5180, (232 1, 38 28) = 5, where for· the· last equality,· · we used the prime· · factorization− from− Exercise 3.2 and the prime factorization 38 28 = (32 22) (32 + 22) (34 + 24) = 5 13 97. − − · · · · Exercise 4.13. We have (2880, 3000, 3240) = (120, 3240) = 120 and [36, 42, 49] = [252, 49] = 1764.
Exercise 4.15. For example, the numbers a1 = 6, a2 = 10, a3 = 15 are rela- tively prime, since (a1, a2, a3) = (6,10,15) = (2,15) = 1. The numbers a1, a2, a3, however, are not pairwise relatively prime, since we have (a1, a2) = 2. N Exercise 4.17. Let a1,..., an . We leave to the reader the proof of the fol- lowing equivalence: ∈
(a ,..., a ) [a ,..., a ] = a a a ,..., a pairwise relatively prime. 1 n · 1 n 1 ··· n ⇐⇒ 1 n This proves the desired criterion.
Exercise 5.2. We obtain 773 = 2 337 + 99. Further, we calculate 25 34 52 = (22 32) (23 32 52) = (5 7 + 1)· (23 32 52) = 7 (23 32 53) + (23· 32· 52). Since· 216· + 1·= 4·8 + 1, it follows· · that· 232 · 1 = (2·16 ·1)(4·8 + 1) + 0.· · − − 252 Solutions to Exercises
Exercise 5.4. This process can be carried out for arbitrary natural numbers g > 1. One obtains the unique representation
` ` 1 1 0 n = q` g + q` 1 g − + + q1 g + q0 g · − · ··· · · with natural numbers 0 q g 1 (j = 0,. . ., `) and q = 0, called the g-adic ≤ j ≤ − ` 6 representation of the natural number n.
Solutions to Exercises in Chapter II
Exercise 1.3. For a, b, c , one has the equalities ∈ Rn (a b) c = (a + b) c = ( (a + b) + c), ⊕ ⊕ Rn ⊕ Rn Rn a (b c) = a (b + c) = (a + (b + c)). ⊕ ⊕ ⊕ Rn Rn Rn Division with remainder of a + b and b + c by n yields the uniquely deter- mined numbers q , q N such that 1 2 ∈ a + b = q n + (a + b) and b + c = q n + (b + c), 1 · Rn 2 · Rn whence follows
( (a + b) + c) = (a + b + c q n) = (a + b + c) Rn Rn Rn − 1 · Rn = (a + b + c q n) = (a + (b + c)). Rn − 2 · Rn Rn We have thereby shown that the operation is associative. Analogously, one can prove that the operation is associative.⊕
Exercise 1.4. (a) The set 2 N = 2 n n N of even natural numbers is a · { · | ∈ } nonempty subset of N. If 2 m, 2 n 2 N, then · · ∈ · 2 m + 2 n = 2 (m + n) 2 N, (2 m) (2 n) = 2 (m 2 n) 2 N. · · · ∈ · · · · · · · ∈ · Thus both + and are operations on 2 N. Since the operations + on N and on N are associative,· it follows that in· particular, the operations + on 2 N and· on 2 N are associative. Therefore, both (2 N, +) and (2 N, ) are· semigroups.· · · · · The set 2 N + 1 = 2 n + 1 n N of odd natural numbers is a · { · | ∈ } nonempty subset of N. If 2 m + 1 and 2 n + 1 are in 2 N + 1, then · · · (2 m + 1) + (2 n + 1) = 2 (m + n + 1) 2 N, · · · ∈ · (2 m + 1) (2 n + 1) = 2 (m 2 n + m + n) + 1 2 N + 1. · · · · · · ∈ · Solutions to Exercises in Chapter II 253
Therefore, while is an operation on 2 N + 1, we see that + is not an opera- tion on 2 N + 1. Therefore,· (2 N + 1, +)· is not a semigroup. The operation on N is associative,· and so in particular,· the operation on 2 N + 1 is asso-· ciative. Therefore, (2 N + 1, ) is a semigroup. · · · · (b) Let k N, k > 1. The set k N = k n n N is a nonempty subset of ∈ · { · | ∈ } N. One shows, as in (a), that both (k N, +) and (k N, ) are semigroups. · · · Exercise 1.5. Because of the inequality
3 3 2 3 2 6 9 2 (2 3) 2 = (2 ) 2 = (2 ) = 2 · = 2 = 2 = 2 (3 ) = 2 (3 2), ◦ ◦ ◦ 6 ◦ ◦ ◦ the operation on N is not associative, and therefore (N, ) is not a semi- group. ◦ ◦
Exercise 1.8. If A = a is a one-element set, then 1 { 1} map(A ) = id , 1 { } where the mapping id : A A is given by the assignment a a . The 1 −→ 1 1 7→ 1 semigroup (map(A ), ) is abelian. If A = a , a ,... is an arbitrary set 1 ◦ 2 { 1 2 } that contains at least two elements a = a , then 1 6 2 map(A ) = id, f , g,... , 2 { } where the mapping id : A A is given by a a (a A ), the mapping 2 −→ 2 j 7→ j j ∈ 2 f : A A by a a , a a (a A a ), and the mapping g : A 2 −→ 2 1 7→ 2 j 7→ j j ∈ 2 \{ 1} 2 −→ A by a a , a a (a A a ). But then we have 2 2 7→ 1 j 7→ j j ∈ 2 \{ 2} ( f g)(a ) = f (g(a )) = f (a ) = a = a = g(a ) = g( f (a )) = (g f )(a ), ◦ 1 1 1 2 6 1 2 1 ◦ 1 whence (map(A ), ) is a nonabelian semigroup. 2 ◦
Exercise 1.12. Let e` be a left identity element and er a right identity element of H. Then e = e e = e , ` ` ◦ r r where the first equality follows from the fact that er is a right identity el- ement of H, and the second from the fact that e` is a left identity element of H.
Exercise 1.14. (a) By Exercise 1.4, (2 N,+) and (2 N, ) are semigroups. It remains to show that there exists an· additive identity· · element in 2 N. By the definition of addition, 0 is this element. Since 1 2 N, there is· no multiplicative identity element, so that (2 N, ) is only a6∈ semigroup.· · · (b) We leave it to the reader to find other examples of semigroups that are not monoids. 254 Solutions to Exercises
Exercise 2.3. (a) Suppose that g0 and g00 are two inverse elements to an ele- ment g G. Then ∈ g0 = g0 e = g0 (g g00) = (g0 g) g00 = e g00 = g00 , ◦ ◦ ◦ ◦ ◦ ◦ where the second equality follows from the fact that g00 is in particular a right inverse to g, and the fourth equality follows from the fact that g0 is in particular a left inverse to g.
(b) Let g`0 be a left inverse and gr0 a right inverse to an element g G. Then it follows that ∈
g0 = g0 e = g0 (g g0 ) = (g0 g) g0 = e g0 = g0 , ` ` ◦ ` ◦ ◦ r ` ◦ ◦ r ◦ r r analogously to part (a).
Exercise 2.6. (a) Let g 1 G be the inverse element to g G. Then − ∈ ∈ 1 1 g g− = e = g− g. ◦ ◦ 1 1 1 Thus g is the inverse element to g− , that is, (g− )− = g. (b) Let g 1 G be the inverse element to g G and h 1 G the inverse − ∈ ∈ − ∈ element to h G. Then ∈ 1 1 1 1 1 1 (h− g− ) (g h) = h− (g− g) h = h− e h = h− h = e, ◦ ◦ ◦ ◦ ◦ ◦ ◦ ◦ ◦ 1 1 1 1 1 1 (g h) (h− g− ) = g (h h− ) g− = g e g− = g g− = e. ◦ ◦ ◦ ◦ ◦ ◦ ◦ ◦ ◦ Thus h 1 g 1 is the inverse element to g h, that is, (g h) 1 = h 1 g 1. − ◦ − ◦ ◦ − − ◦ − The calculational rules (c) and (d) follow directly from the definition.
Exercise 2.9. We compare only the groups that have the same numbers of el- ements. The Cayley tables of the groups ( 4, ), ( 5 0 , ), and (D4, ) have, reading from left to right, the followingR ⊕ form:R \{ } ◦
0 1 2 3 1 2 3 4 d d s s ⊕ ◦ 0 1 0 1 0 0 1 2 3 1 1 2 3 4 d0 d0 d1 s0 s1 1 1 2 3 0 2 2 4 1 3 d1 d1 d0 s1 s0 2 2 3 0 1 3 3 1 4 2 s0 s0 s1 d0 d1 3 3 0 1 2 4 4 3 2 1 s1 s1 s0 d1 d0 One may conclude from the Cayley tables that all three groups under con- sideration are abelian. We now determine the smallest nonzero natural num- ber n such that gn = e for g = e. In the group ( , ), we have e = 0 and 6 R4 ⊕ 12 = 2, 13 = 3, 14 = 0; 22 = 0; 32 = 2, 33 = 1, 34 = 0.
In the group ( 0 , ), we have e = 1 and R5 \{ } Solutions to Exercises in Chapter II 255
22 = 4, 23 = 3, 24 = 1; 32 = 4, 33 = 2, 34 = 1; 42 = 1.
Thus in each group there are two elements with n = 4 and one element with n = 2. In the group (D , ), however, d2 = s2 = s2 = e with e = d , that is, 4 ◦ 1 0 1 0 there is no element with n = 4. The Cayley tables for ( 6, ) and (D6, ), reading from left to right, have the following form: R ⊕ ◦
0 1 2 3 4 5 d d d s s s ⊕ ◦ 0 1 2 0 1 2 0 0 1 2 3 4 5 d0 d0 d1 d2 s0 s1 s2 1 1 2 3 4 5 0 d1 d1 d2 d0 s2 s0 s1 2 2 3 4 5 0 1 d2 d2 d0 d1 s1 s2 s0 3 3 4 5 0 1 2 s0 s0 s1 s2 d0 d1 d2 4 4 5 0 1 2 3 s1 s1 s2 s0 d2 d0 d1 5 5 0 1 2 3 4 s2 s2 s0 s1 d1 d2 d0 From these tables, one can see that the group ( , ) is abelian. The group R6 ⊕ (D , ) is nonabelian, since s s = d = d = s s . We again determine 6 ◦ 0 ◦ 1 2 6 1 1 ◦ 0 the smallest nonzero natural number n such that gn = e for g = e. In ( , ), 6 R6 ⊕ we have e = 0 and
12 = 2, 13 = 3, 14 = 4, 15 = 5, 16 = 0; 22 = 4, 23 = 0; 32 = 0; 42 = 2, 43 = 0; 52 = 4, 53 = 3, 54 = 2, 55 = 1, 56 = 0.
There exist, therefore, in ( , ) two elements with n = 6, two elements R6 ⊕ with n = 3, and one element with n = 2. In (D , ), we have e = d and 6 ◦ 0 2 3 2 3 2 2 2 d1 = d2, d1 = d0; d2 = d1, d2 = d0; s0 = d0, s1 = d0, s2 = d0.
There exist, therefore, in (D , ) no element with n = 6, two elements with 6 ◦ n = 3, and three elements with n = 2.
Exercise 2.10. (a) One shows using the Cayley table
1 2
1 1 2 2 2 1 for ( 3 0 , ) and the Cayley table from Exercise 2.9 for ( 5 0 , ) that (R \{ 0} , ) and ( 0 , ) are groups. R \{ } R3 \{ } R5 \{ } (b) We leave to the reader the task of verifying the assertions of Exam- ple 2.8 (iii) regarding the dihedral group (D , ). 2n ◦ (c) Let n 3. We consider the elements ≥ 1 2 3 n 1 2 3 n π = ··· , π = ··· 1 2 1 3 n 2 3 1 2 n ··· ··· 256 Solutions to Exercises of Sn, where for n > 3, each of the elements 4,. . .,n is mapped to itself. Then 1 2 3 n 1 2 3 n π π = ··· = ··· = π π , 1 ◦ 2 3 2 1 n 6 1 3 2 n 2 ◦ 1 ··· ··· where for n > 3, each of the elements 4,. . .,n is mapped to itself. This proves that (S , ) for n 3 is nonabelian. n ◦ ≥ Exercise 2.13. This is proved by induction on n.
Exercise 2.19. We have S = π , π , π , π , π , π with 3 { 1 2 3 4 5 6} 1 2 3 1 2 3 1 2 3 π = , π = , π = , 1 1 2 3 2 2 3 1 3 3 1 2 1 2 3 1 2 3 1 2 3 π = , π = , π = . 4 1 3 2 5 3 2 1 6 2 1 3
We calculate ord(π1) = 1, ord(π2) = ord(π3) = 3, and ord(π4) = ord(π5) = ord(π6) = 2.
Exercise 2.23. Since dk = d (k = 0,. . .,n 1) and dn = d , we have d = 1 k − 1 0 h 1i d0,..., dn 1 . Using the subgroup criterion, one can show that the nonempty { − } subset g = ..., (g 1)2, g 1, g0 = e, g1 = g, g2,... G is a subgroup of G h i { − − } ⊆ for every group G. In particular, d1 = d0,..., dn 1 is a subgroup of D2n. h i { − }
Exercise 2.26. We have S3 = π1, π2, π3, π4, π5, π6 with πj (j = 1,. . .,n) as in Exercise 2.19. We have the{ cyclic subgroups }
π = π , π = π ,π ,π = π , h 1i { 1} h 2i { 1 2 3} h 3i π = π ,π , π = π ,π , π = π ,π , h 4i { 1 4} h 5i { 1 5} h 6i { 1 6} and the subgroup S3 itself, which is not cyclic. One can see that S3 has no other subgroups.
Exercise 3.3. We have S = π , π , π , π , π , π with π (j = 1,. . .,n) as 3 { 1 2 3 4 5 6} j in Exercise 2.19 and D6 = d0, d1, d2, d0 s0, d1 s0, d2 s0 , where s0 is re- flection in the median of the{ side joining◦ vertices◦ 1 and◦ 2.} By the definition of the group homomorphism f : D S , we have 6 −→ 3
f (d0) = π1, f (d1) = π3, f (d2) = π2, f (d s ) = π , f (d s ) = π , f (d s ) = π , 0 ◦ 0 6 1 ◦ 0 5 2 ◦ 0 4 which proves that f is bijective and therefore in fact a group isomorphism.
k1 k2 Exercise 3.5. Let dj1 s0 and dj2 s0 with j1, j2 0,. . ., n 1 and k1, k2 ◦ ◦ ∈ { 1 − } ∈ 0, 1 be two elements of D . Since d s = s d− , we have { } 2n j ◦ 0 0 ◦ j Solutions to Exercises in Chapter II 257 d d , if k = 0, k = 0; j1 j2 1 2 ◦ d d s0, if k1 = 0, k2 = 1; k1 k2 j1 ◦ j2 ◦ (dj1 s0 ) (dj2 s0 ) = 1 ◦ ◦ ◦ dj d− , if k1 = 1, k2 = 1; 1 ◦ j2 1 dj d− s0, if k1 = 1,k2 = 0. 1 ◦ j2 ◦ It follows that