Section 10 Cosets and the Theorem of Lagrange 97

Theory
46. Prove that a direct product of abelian groups is abelian.
47. Let G be an abelian group. Let H be the subset of G consisting of the identity e together with all elements of
G of order 2. Show that H is a subgroup of G.
48. Following up the idea of Exercise 47 determine whether H will always be a subgroup for every abelian group
G if H consists of the identity e together with all elements of G of order 3; of order 4. For what positive
integers n will H always be a subgroup for every abelian group G, if H consists of the identity e together with
all elements of G of order n? Compare with Exercise 54 of Section 5.
49. Find a counterexample of Exercise 47 with the hypothesis that G is abelian omitted.
Let H and K be subgroups of a group G. Exercises 50 and 51 ask you to establish necessary and sufficient criteria
for G to appear as the internal direct product of H and K.
50. Let H and K be groups and let G = H x K. Recall that both H and K appear as subgroups of G in a natural
way. Show that these subgroups H (actually H x {e}) and K (actually {e} x K) have the following properties.
a. Every element of G is of the form hk for some h E H and k E K .
b. hk = kh for all h E H and k EK. c. H n K = {e}.

51. Let H and K be subgroups of a group G satisfying the three properties listed in the preceding exercise. Show
that for each g E G, the expression g = hk for h E H and k E K is unique. Then let each g be renamed (h, k).
Show that, under this renaming, G becomes structurally identical (isomorphic) to H x K.
52. Show that a finite abelian group is not cyclic if and only if it contains a subgroup isomorphic to Zp x Zp for
some prime p.
53. Prove that if a finite abelian group has order a power of a prime p, then the order of every element in the group
is a power of p.
54. Let G, H , and K be finitely generated abelian groups. Show that if G x K is isomorphic to H x K, then G '.:::'.H.
55. Using the notation of Theorem 9.14, prove that for any finite abelian group G, every cyclic subgroup of G has
order no more than dk> the largest invariant factor for G.

SECTION 10 COSETS AND THE THEOREM OF LAGRANGE

You may have noticed that the order of a subgroup H of a finite group G seems always
to be a divisor of the order of G. This is the theorem of Lagrange. We shall prove it by
exhibiting a partition of G into cells, all having the same size as H. Thus if there are r
such cells, we will have
r(order of H ) = (order of G)
from which the theorem follows immediately. The cells in the partition will be called
cosets of H , and they are important in their own right. In Section 12, we will see that if
H satisfies a certain property, then each coset can be regarded as an element of a group
in a very natural way. We give some indication of such coset groups in this section to
help you develop a feel for the topic.

Co sets
Let H be a subgroup of a group G, which may be of finite or infinite order. We exhibit a
partition a G by defining an equivalence relation, ~Lon G.

10.1 Theorem Let H be a subgroup of G. Let the relation ~L be defined on G by

a~Lb if and only if a- 1bEH.

Then ~L i s an equivalence relation on G.

98 Part II Structure of Groups

Proof When reading the proof, notice how we must constantly make use of the fact that H is a
subgroup of G.

Reflexive Let a E G. Then a - 1a = e and e E H since His a subgroup. Thus

a ~ La .
Symmetric Suppose a ~Lb. Then a- 1b E H. Since His a subgroup, (a- 1b)- 1
is in Hand (a- 1b)- 1 = b- 1a, so b- 1a is in Hand b ~La.
Transitive Let a ~ L b and b ~L c. Then a- 1b E Hand b- 1c E H . Since H is
a subgroup, (a- 1b)(b- 1c) = a- 1c is in H , so a ~L c. +
The equivalence relation ~ L in Theorem 10.1 defines a partition of G, as described
in Theorem 0.22. Let's see what the cells in this partition look like. Suppose a E G. The
cell containing a consists of all x E G such that a ~L x, which means all x E G such
that a- 1x E H. Now a- 1x EH if and only if a- 1x = h for some h EH, or equivalently,
if and only if x = ah for some h E H. Therefore the cell containing a is {ah I h E H},
which we denote by aH.

10.2 Definition Let H be a subgroup of a group G. The subset aH = {ah I h EH) of G is the left coset
of H containing a. •

10.3 Example Exhibit the left coset of the subgroup 3Z of Z.

Solution Our notation here is additive, so the left coset of 3Z containing m is m + 3Z. Taking
m = 0, we see that

3Z = {· · · , -9, -6, -3,0,3,6, 9, · · ·}

is itself one of its left cosets, the coset containing 0. To find another left coset, we select
an element of Z not in 3Z, say 1, and find the left coset containing it. We have

1+3Z = {· · ·, -8, -5, -2, 1, 4, 7, 10, · · · }.

These two left cosets, 3Z and I + 3Z, do not yet exhaust Z . For example, 2 is in neither
of them. The left coset containing 2 is

2 + 3Z = {· · ·, -7, -4, - 1,2, 5, 8, 11, · · · }.

It is clear that these three left cosets we have found do exhaust Z, so they constitute the
partition of Z into left cosets of 3Z. .&.

10.4 Example We find the partition of Z 12 into left cosets of H = (3). One coset is always the subgroup
itself, so 0 + H = {O, 3, 6, 9). We next find I + H = {I , 4, 7, 10). We are still not done
since we have not included every element of Z 12 in the two cosets we listed so far.
Finally, 2 + H = {2, 5, 8, 11) and we have computed all the left cosets of Hin Z12 . .&.

10.5 Example We now list the left cosets of the subgroup H = (µ,) = {t, µ,) of the nonabelian group
n4 = {t, p, p2, p 3, µ,, µ,p , µ,p2, µ,p 3).
l{l, µ,) {l , µ,)
p{t , µ,) {p, µ,p 3)
p2 {t ,µ, ) {p2, µ,p2)
p 3{l , µ,) {p3,µ,p)

We know this is a complete list of the left cosets since every element of D4 appears in
exactly one of the listed sets. .&.
 Section 10 Cosets and the Theorem of Lagrange 99

The Theorem of Lagrange

In Example 10.4 each left coset of (3) ::: :Z. 12 has four elements. In Example 10.5, each
left coset has two elements. From the computation of the left cosets, it is no surprise that
all left cosets of a subgroup have the same number of elements. Theorem 10.6 confirms
this is what happens in general.

10.6 Theorem Let H be a subgroup of G. Then for any a E G, the coset aH has the same cardinality
as H.

Proof Let f : H ~ aH be defined by the formula f(h) = ah. To show f is one-to-one, we

suppose that b, c EH andf(b) = f(c). Then ab = ac and left cancellation gives b =c.
So f is one-to-one. Now suppose that y E aH. Then there is an h E H such that y = ah
by definition of the left coset aH. Thus y = f(h) and f is surj ective. Since there is a
one-to-one function mapping H onto aH , Hand aH have the same cardinality. +
In the case of a finite subgroup H , Theorem 10.6 says that H and aH have the same
number of elements for any a in the group G. This is precisely what we were seeking in
order to prove Lagrange's Theorem.

10.7 Theorem (Theorem of Lagrange) Let H be a subgroup of a finite group G. Then the order of
Hi s a divisor of the order of G.

Proof Let n be the order of G, and let H have order m. Theorem 10.6 shows that every coset of
H also has m elements. Let r be the number of cells in the partition of G into left cosets
of H. Then n = rm, so m is indeed a divisor of n. +
Note that this elegant and important theorem comes from the simple counting of
cosets and the number of elements in each coset. We continue to derive consequences
of Theorem 10.7, which should be regarded as a counting theorem.

10.8 Corollary Every group of prime order is cyclic.

Proof Let G be of prime order p, and let a be an element of G different from the identity. Then
the cyclic subgroup (a) of G generated by a has at least two elements, a and e. But
by Theorem 10.7, the order m '.'.: 2 of (a) must divide the prime p. Thus we must have
m = p and (a) = G, so G is cyclic. +
Since every cyclic group of order p is isomorphic to :Z.1" we see that there is only
one group structure, up to isomorphism, of a given p rime order p. Now doesn' t this
elegant result follow easily from the theorem of Lagrange, a counting theorem? Never
underestimate a theorem that counts something. Proving the preceding corollary is a
favorite examination question.

10.9 Theorem The order of an element of a finite group divides the order of the group.

Proof Remembering that the order of an element is the same as the order of the cyclic subgroup
generated by the element, we see that this theorem follows directly from Lagrange's
Theorem. +
10.10 Definition Let H be a subgroup of a group G. The number of left cosets of H in G is the index
(G : H) of Hin G. •
The index (G: H) j ust defined may be finite or infinite. If G is finite, then obviously
(G : H) is finite and (G : H) = IGl / IH I, since every coset of H contains IH I elements.
We state a basic theorem concerning indices of subgroups, and leave the proof to the
exercises (see Exercise 40).
100 Part II Structure of Groups

10.11 Theorem Suppose Hand K are subgroups of a group G such that K :S H :S G, and suppose (H: K )
and (G: H) are both finite. Then (G: K) is finite, and (G: K) = (G: H)(H: K) .

Lagrange's Theorem says that for any subgroup H of a finite group G, the order
of H divides the order of G. But if d is a divisor of the order of G, does G necessarily
have a subgroup with exactly d elements? We will show in Section 13 that the answer is
no for some groups. This suggests a new question: Under what conditions does G have
a subgroup of every order d that is a divisor of G? We saw in Section 9 that for every
divisor of the order of an abelian group, there is a subgroup of that order. The complete
answer to this question is beyond the scope of this book, but we will come back to the
question later.

Cosets Left and Right!

It is possible to do everything we have done in this section using right cosets instead of
left cosets. All it takes is some minor and straightforward modifications to the defini-
tions and proofs. We briefly give the corresponding definitions that lead to right cosets
and point out some of their properties.
Let H be a subgroup of G. To start with, instead of ~L we could have used ~R
defined by
a ~Rb if and only if ab- 1 E H.
With this definition, ~R is an equivalence relation and the equivalence classes are the
right cosets. The right coset of H containing the element a E G is

Ha= {ha f h EH}.

Just like left cosets, each right coset of a subgroup H has the same cardinality as H. So
left cosets and right cosets have the same cardinality. In abelian groups, the right and
left cosets are the same, but there is no reason to think they would be the same in general
for nonabelian groups. If the right and left cosets are the same, we can drop left or right
and just refer to cosets.

10.12 Example In Example 10.5 we computed the left cosets of the subgroup H = (µ,) = {i, µ,}of the
group D4 = {i , p , p 2, p 3, µ , µp , µp 2, µp 3}. We now compute the right cosets.

{t , µ}l {i , µ}
{t , µ}p {p , µp }
{t, µ} p2 {p2, µp2}
{t, µ}p3 {p3' µp 3}

The right cosets and the left cosets are not the same. For example, p H = {p, µp 3} while
Hp = {p,µp}. .&.

If this were the whole story of left and right cosets, there would be no reason to even
mention right cosets. We could just use left coset, prove Lagrange's Theorem, and call
it a day. However, as we shall see in Part III, a curious thjng happens when the left and
right cosets are the same. We illustrate with an example.

10.13 Example The group &::6 is abelian. Find the partition of &::6 into cosets of the subgroup H = (0, 3}.

Solution One coset is {0, 3} itself. The coset containing 1 is 1 + (0, 3} = {1, 4}. The coset con-
taining 2 is 2 + {O, 3} = {2, 5}. Since {O, 3}, {l , 4}, and {2, 5} exhaust all of &::6, these
are all the cosets. .&.
 Section 10 Cosets a nd t he Theorem of Lagrange 101

We point out a fascinating thing that we will develop in detail in Section 12. Re-
ferring back to Example 10.13, Table 10.14 gives the binary operation for Z6 but with
elements listed in the order they appear in the cosets (0, 3), {1, 4), (2, 5). We shaded the
table according to these cosets.

10.14 Table 10.15 Table

+6 0 3 1 4 2 5
0 0 3 I 4 2 5 LT MD DK
3 3 0 4 1 5 2
LT LT MD DK
1 1 4 2 5 3 0
4 4 1 5 2 0 3 MD MD DK LT
12 2 5 3 0 4 1
15 5 2 0 3 1 4 DK DK LT MD

Suppose we denote these cosets by LT(light), MD(medium), and DK(dark) ac-

cording to their shading. Table 10.14 then defines a binary operation on these shadings,
as shown in Table 10.15. Note that if we replace LT by 0, MD by 1, and DK by 2 in
Table 10.15, we obtain the table for Z 3 . Thus the table of shadings forms a group!
We will see in Section 12 that when left cosets and right cosets are the same, then
the cosets form a group as in Example 10.13. If right and left cosets are different, the
construction fails.

10.16 E xample Let H = {t, µ,) '.'S D 3 . The group table for D 3 is given below with the elements arranged
so that left cosets are together. The double lines divide the cosets.

The situation here is much different from the situation in Example 10.13. In Table 10.14
the two-by-two blocks in the table each contain only elements of a left coset. In the
present example, most blocks do not contain elements from only one left coset. Further-
more, even if we tried to use the two-by-two blocks of elements to form a three-by-three
group table, the second row of blocks contains two blocks, both having the same ele-
ments, {p 2 , µ,p , µ,, t). So the table of blocks would have a row with the same element
listed twice. In this case, there is no natural way of making the left cosets a group. .A

If G is an abelian group, then the left and right cosets are the same. Theorem 10.17
gives another condition when left and right cosets are the same. Recall that if if> : G --+
G' is a group homomorphism, then Ker(ef>) = ef>- 1[{e }] '.'S G is the kernel of if>.

10.17 Theor em Let if> : G--+ G' be a group homomorphism. Then the left and right cosets of Ker(ef>)
are identical. Furthermore, a, b E G are in the same coset of Ker(ef>) if and only if
ef>(a) = ef>(b).
102 Part II Structure of Groups

Proof We first assume that a and bare in the same left cosets of Ker(<f>) and show they are also
in the same right cosets. Then a- 1b E Ker(<f>). So <f>(a- 1b) = e, the identity element.
Because <t> is a homomorphism, <f>(a)- 1</>(b) = e, which implies that <f>(a) = <f>(b).
Therefore, <f>(ab- 1) = <f>(a)<f>(b)- 1 = <f>(a)<f>(a)- 1 = e. Thus ab- 1 E Ker(<f>), which says
that a and b are in the same right coset. Note that in the process we showed that if a and
bare in the same left coset of Ker(<f> ), then <f>(a) = <f>(b).
Now suppose that <f>(a) = <f>(b). Then <f>(b- 1a) = <f>(b)- 1<f>(a) =e. Thus b- 1a E
Ker(<t> ), which implies that a and b are in the same left coset.
To complete the proof, we need to show that if a and b are in the same right coset,
then they are also in the same left coset. The proof is essentially the same as above, so
we leave this detail to the reader. +
10.18 Example Consider the determinant map det : GL(2, JR) --+ JR*. In linear algebra you learn that
det(AB) = det(A) det(B), so the determinant is a group homomorphism. The kernel of
det is the set of all 2 x 2 matrices with determinant L Two matrices A , B E GL(2, JR)
are in the same left coset of Ker(det) if and only if they are in the same right coset of
Ker(det) if and only if det(A) = det(B). In particular, the two matrices

[ ~ ~ J and [ ~ ; J
each have determinant 2, so they are in the same left (and right) cosets of Ker(det). .A.

10.19 Corollary A homomorphism <t> : G--+ G' is o ne-to-one if and only if Ker(</>) is the trivial subgroup
of G.

Proof We first assume that Ker(<t>) = {e}. Every coset of Ker(<t>) has only one element. Suppose
that <f>(a) = <f>(b). Then a and b are in the same coset of Ker(<f>) by Theorem 10.17. Thus
a= b.
Now suppose that <t> is one-to-one. Then only the identity e is mapped to the identity
in G' . So Ker(</>)= {e}. +
Corollary 10.19 says that to check if a homomorphism <t> : G --+ G' is one-to-one
one merely needs to check that Ker(</>) is the trivial subgroup. In other words, show
that the only solution to <f>(x) = e' is e, where e and e' are the identities in G and G' ,
respectively.

10.20 Example Let <f> : lR--+ JR+ be defined by <f>(x) = 2x_Since <f> is a homomorphism, we can check
that <t> is one-to-one by solving <f>(x) = 1. The equation 2x = <f>(x) = 1 has only the
solution 0 since for x > 0, 2x > 1 and for x < 0, 2x < 1. Thus <f> is one-to-one. .A.

• EXERCISES 10

Computations
1. Find all cosets of the subgroup 41:: of Z .
2. Find all cosets of the subgroup 41:: of 21::.
3. Find all cosets of the subgroup (3) in Z is-
4. Find all cosets of the subgroup (6) in Z 1s-
S. Find all cosets of the subgroup ( 18) of 1::36·
6. Find all left cosets of (µp) in D4-
7. Repeat the preceding exercise, but find the right cosets this time. Are they the same as the left cosets?
 Section 10 Exercises 103

8. Are the left and right cosets the same for the subgroup {t, p 4 , µ.,, µ.,p 4 } of Ds? If so, display the cosets. If not,
find a left coset that is not the same as any right coset.
9. Find all the left cosets of (p2 ) :::= D4.
10. Repeat the previous exercise, but find the right cosets. Are the left and right cosets the same? If so, make the
group table for D4, ordering the elements so that the cosets are in blocks, see if the blocks form a group with
four elements, and determine what group of order 4 the blocks form.
11. Find the index of (p2 ) in the group D6.
12. Find the index of (3) in the group Z24.
13. Find the index of l 2Z in Z.
14. Find the index of 12Z in 3Z.
15. Let a = (1 , 2 , 5, 4)(2, 3) in S5. Find the index of (a ) in S5.
16. Letµ., = (I, 2, 4, 5)(3, 6) in 56. Find the index of (µ.,) in 56.

Concepts
In Exercises 17 through 19, correct the definition of the italicized term without reference to the text, if correction
is needed, so that it is in a form acceptable for publication.

17. Let G be a group and let H ~ G. The left coset of H containing a is aH ={ah I h EH}.
18. Let G be a group and let H :::= G. The index of H in G is the number of right cosets of Hin G.
19. Let 4>: G--+ G'. Then the kernel of 4> is Ker(</>)= {g E G I ef>(g) = e) .
20. Determine whether each of the following is true or false.

a. Every subgroup of every group has left cosets.

b. The number of left cosets of a subgroup of a finite group divides the order of the group.
c. Every group of prime order is abelian.
d. One cannot have left cosets of a finite subgroup of an infinite group.
e. A subgroup of a group is a left coset of itself.
f. Only subgroups of finite groups can have left cosets.
g. A11 is of index 2 in S11 for n > 1.
h. The theorem of Lagrange is a nice result.
i. Every finite group contains an element of every order that divides the order of the group.
j. Every finite cyclic group contains an element of every order that divides the order of the group.
k. The kernel of a homomorphism is a subgroup of the range of the homomorphism.
I. Left cosets and right cosets of the kernel of a homomorphism are the same.

In Exercises 2 1 through 26, give an example of the desired subgroup and group if possible. If impossible, say why
it is impossible.

21. A subgroup H :::= G with G infinite and H having only a finite number of left cosets in G
22. A subgroup of an abelian group G whose left cosets and right cosets give different partitions of G
23. A subgroup of a group G whose left cosets give a partition of G into just one cell
24. A subgroup of a group of order 6 whose left cosets give a partition of the group into 6 cells
25. A subgroup of a group of order 6 whose left cosets give a partition of the group into 12 cells
26. A subgroup of a group of order 6 whose left cosets give a partition of the group into 4 cells

Proof Synopsis
27. Give a one-sentence synopsis of the proof of the Theorem of Lagrange.
104 Part II Structure of Groups

Theory
28. Prove that the relation ~R that is used to define right cosets is an equivalence relation.
29. Let H be a subgroup of a group G and let g E G. Define a one-to-one map of H onto Hg. Prove that your map
is one-to-one and is onto Hg.
30. Let H be a subgroup of a group G such that g- 1hg E H for all g E G and all h E H . Show that every left coset
gH is the same as the right coset Hg.
31. Let H be a subgroup of a group G. Prove that if the partition of G into left cosets of H is the same as the
partition into right cosets of H, then g- 1hg E H for all g E G and all h E H. (Note that this is the converse of
Exercise 30.)
Let H be a subgroup of a group G and let a, b E G. In Exercises 32 through 35 prove the statement or give a
counterexample.
32. If aH = bH, then Ha = Hb.
33. If Ha = Hb, then b E Ha.
34. If aH = bH, then Ha - 1 = H b- 1•

35. If aH = bH , then a 2
H = b H.2

36. Let G be a group of order pq, where p and q are prime numbers. Show that every proper subgroup of G is
cyclic.
37. Show that there are the same number of left as right cosets of a subgroup H of a group G; that is, exhibit
a one-to-one map of the collection of left cosets onto the collection of right cosets. (Note that this result is
obvious by counting for finite groups. Your proof must hold for any group.)
38. Exercise 29 of Section 2 showed that every finite group of even order 2n contains an element of order 2.
Using the theorem of Lagrange, show that if n is odd, then an abelian group of order 2n contains precisely one
element of order 2.
39. Show that a group with at least two elements but with no proper nontrivial subgroups must be finite and of
prime order.
40. Prove Theorem 10.11 [Hint: Let {a;H Ii = 1, · · · , r ) be the collection of distinct left cosets of H in G and
{bj K Ij = 1, · · · , s} be the collection of distinct left cosets of K in H . Show that

{(a;bj)K f i= 1,- ·· ,r; j= 1,- ·· , s)

is the collection of distinct left cosets of K in G.]

41. Show that if His a subgroup of index 2 in a finite group G, then every left coset of His also a right coset of H.
42. Show that if a group G with identity e has fi nite order n, then an = e for all a E G.
43. Show that every left coset of the subgroup Z of the additive group of real numbers contains exactly one element
x such that 0 :'.": x < 1.
44. Show that the function sine assigns the same value to each element of any fixed left coset of the subgroup (2rr )
of the additive group lR of real numbers. (Thus sine induces a well-defined function on the set of cosets; the
value of the function on a coset is obtained when we choose an element x of the coset and compute sin x.)
45. Let H and K be subgroups of a group G. Define ~ on G by a ~ b if and only if a = hbk for some h E H and
some k E K.
a. Prove that ~ is an equivalence relation on G.
b. Describe the elements in the equivalence class containing a E G. (These equivalence classes are called
double cosets.)
46. Let SA be the group of all permutations of the set A , and let c be one particular element of A.
a. Show that {a E SA I a(c) = c} is a subgroup Sc,c of SA .
b. Let d =I- c be another particular element of A. Is Sc,d = {a E SA I a(c) = d ) a subgroup of SA? Why or why
not?
c. Characterize the set Sc,d of part (b) in terms of the subgroup Sc,c of part (a).
 Section 11 Plane Isometries 105

47. Show that a finite cyclic group of order n has exactly one subgroup of each order d dividing n, and that these
are all the subgroups it has.
48. The Euler phi-function is defined for positive integers n by rp(n) = s, where s is the number of positive
integers less than or equal ton that are relatively prime ton. Use Exercise 47 to show that

n = I : rp<d).
d in

the sum being taken over all positive integers d dividing n. [Hint: Note that the number of generators of Z d is
rp(d) by Corollary 6. 17.]
49. Let G be a finite group. Show that if for each positive integer m the number of solutions x of the equation
x" = e in G is at most m, then G is cyclic. [Hint: Use Theorem 10.9 and Exercise 48 to show that G must
contain an element of order n = IGI.]
50. Show that a finite group cannot be written as the union of two of its proper subgroups. Does the statement
remain true if "two" is replaced by "three"? (This was problem B-2 on the 1969 Putnam Exam.)

SECTION 11 t pLANE ISOMETRIES

Consider the Euclidean plane IR2 . An isometry of IR2 is a permutation </J : IR2 ---+ IR2
that preserves distance, so that the distance between points P and Q is the same as
the distance between the points </J(P) and </J(Q) for all points P and Q in IR2 . If 1/1 is
also an isometry of IR2 , then the distance between 1/1(</J(P)) and 1/1(</J(Q)) must be the
same as the distance between </J(P) and </J(Q), which in tum is the distance between P
and Q, showing that the composition of two isometries is again an isometry. Since the
identity map is an isometry and the inverse of an isometry is an isometry, we see that
the isometries of IR2 form a subgroup of the group of all permutations of IR2 .
Given any subset S of IR2 , the isometries of IR2 that carry S onto itself form a
subgroup of the group of isometries. This subgroup is the group of symmetries of Sin
IR2 . Although we defined the dihedral group D 11 as one-to-one maps from the vertices
of a regular n-gon onto itself that preserves edges, we can extend each map in D 11 to an
isometry of the whole plane; µ, is reflection across the x-axis and p is rotation about the
origin by Zrr.
11
So we can think of D11 as the group of isometries of a regular n-gon in IR2 .
Everything we have defined in the two preceding paragraphs could equally well
have been done for n-dimensional Euclidean space IR", but we will concern ourselves
chiefly with plane isometries here.
It can be proved that every isometry of the plane is one of just four types (see Artin
[5]). We will list the types and show, for each type, a labeled figure that can be carried
into itself by an isometry of that type. In each of Figs. 11.1, 11.3, and 11.4, consider the
line with spikes shown to be extended infinitely to the left and to the right. We also give
an example of each type in terms of coordinates.
translation r: Slide every point the same distance in the same direction. See
Fig. 11. 1. (Example: r(x,y) = (x, y) + (2, -3) = (x + 2,y - 3).)
rotation p: Rotate the plane about a point P through an angle e. See Fig. 11.2.
(Example: p(x,y) = (-y,x) is a rotation through 90° counterclockwise about the
origin (0, 0).)
reflection µ,: Map each point into its mirror image (µ, for mirror) across a line
L, each point of which is left fixed by µ. See Fig. 11.3. The line L is the axis of
reflection. (Example: µ,(x,y) = (y,x) is a reflection across the line y = x.)

t This section is not used in the remainder of the text.