Solutions to Chapter 11 Exercises

in Discrete Mathematics by Norman L. Biggs;

2nd Edition 2002

11.1 Binomial Numbers

n


11.1.1 Show that r
= 0 if r > n.

Discussion As in the text, we regard nr as the binomial number

which measures the size of the set of r -subsets of an n-set.

Solution If r > n then an n-set has no r-subsets.

1
 Solutions to Chapter 11 Exercises
in Discrete Mathematics by Norman L. Biggs;
2nd Edition 2002

n n n




11.1.2 Find the values of 0
, 1
, n
for all n ≥ 1.
n


Solution 0
is the number of empty subsets of an n-set, which is
clearly 1.

n


1
is the number of 1-subsets (‘singletons’) of an n-set, which
is clearly n.

n


n
is the number of n-subsets of an n-set, which is clearly 1.

2
 Solutions to Chapter 11 Exercises
in Discrete Mathematics by Norman L. Biggs;
2nd Edition 2002

n n



11.1.3 Prove that r
= n−r
for 0 ≤ r ≤ n.

Discussion We use the approach mentioned in Discussion of Exercise 1

and the results from Exercise 2. We assume that n ≥ 1. The nota-
tion T ∈ X was discussed at Exercise 2.3.3.

Solution If 0 ≤ r ≤ n, let X and Y respectively denote the

sets of r-subsets and of (n − r)-subsets of some n-set S. If T ∈ X
define f (T ) = S \ T = {s ∈ S | s ∈ / T }. It is clear that f is a
function defined for all elements T ∈ X, and that each f (T ) has
n − r elements, so that f (T ) ∈ Y . It is also clear that f is injective,
and that if U ∈ Y then U = f ((S \ U )), so that f is a bijection.
Hence nr = |X| = |Y | = n−r n



.

3
 Solutions to Chapter 11 Exercises
in Discrete Mathematics by Norman L. Biggs;
2nd Edition 2002

:
11.1.4 Calculate the next three rows of Pascal’s triangle, starting
from the portion given on p. 106.

Discussion We give the full triangle, with the additional rows shown
in bold:

Solution The table is:

1
1 1
1 2 1
1 3 3 1
1 4 6 4 1
1 5 10 10 5 1
1 6 15 20 15 6 1
1
7 21 35 35 21 7 1
1 8
28 56 70 56 28 8 1
1 9 36 84 126 126 84 36 9 1
1 10 45 120 210 252 210 120 45 10 1

4
 Solutions to Chapter 11 Exercises
in Discrete Mathematics by Norman L. Biggs;
2nd Edition 2002

16 17



11.1.5 Evaluate 4
and 5
.

Solution From Theorem 11.1.2,

16


4
= 16!/((12!) × (4!) = (16 × 15 × 14 × 13)/(4 × 3 × 2 × 1) =
2 × 5 × 14 × 13 = 1820.
and
17 16



5
= 17!/((12!) × (5!) = (17 × 4
)/5 = 1820 × 17/5 = 6188.

5
 Solutions to Chapter 11 Exercises
in Discrete Mathematics by Norman L. Biggs;
2nd Edition 2002

11.1.6 Show that the number of words of
length n in the alphabet

{0, 1} which contain exactly r zeros is nr .

Solution The set of ‘positions’ for a character is an n-set, e.g. if

we label them ‘first from left’, ‘second from left’, . . .. Each word
is completely determined by the set of positions in which the 0’s
appear, and different such sets give different words, so there is a
bijection between the set of words and the set of r-subsets
of the set
n
of positions. We know the number of those subsets is r , so that is
also the number of words of length n with exactly r 0’s.

6
 Solutions to Chapter 11 Exercises
in Discrete Mathematics by Norman L. Biggs;
2nd Edition 2002

11.2 Unordered Selections with Repetition

11.2.1 Write down the values given by the table above when r = 2
and n = 3, and in each of the four cases make a list of the relevant
selections, taking {a, b, c} as the 3-set.

Discussion The table in question is Table 11.2.1, and we let X =

{a, b, c}. Note that an unordered selection without repetition is a
set while an ordered selection is a sequence.

Solution It is straightforward to substitute into the formulae from

Table 11.2.1 to find that
• there are 3 × . . . × (3 − 2 + 1) = 6 ordered selections, without
repetition, of 2 elements of X; they are: ab, ba, ac, ca, bc, cb.
• there are 32 = 9 ordered selections, with repetition, of 2
elements of X; they are: aa, bb, cc, ab, ba, ac, ca, bc, cb.
there are 32 = 3!/((2!) × (1!)) = 3 unordered selections,


•
without repetition, of 2 elements of X; they are: {a, b}, {a, c}, {b, c}.
• there are 3+2−1


2
= 4!/((2!)×(2!)) = 6 unordered selections,
with repetition, of 2 elements of X; they are: aa, ab, ac, bb, bc, cc.

10
 Solutions to Chapter 11 Exercises
in Discrete Mathematics by Norman L. Biggs;
2nd Edition 2002

11.2.2 Show that when three indistinguishable dice are thrown

there are 56 possible outcomes. What is the number of outcomes
when n indistinguishable dice are thrown?

Discussion This problem is ‘small’ enough to be solved by consid-

ering three easily recognised cases. But the important general point
is to recognise which of the formulae in the section is appropriate to
the given problem.

Solution The possible outcomes can be partitioned into the cases

when all three values are the same, of which there are 6, those when
one value is repeated exactly twice, of which there are 6 × 5 =
30,
and those when all three values are distinct, when there are 63 =
20, giving a total of 56
More briefly, we are interested in the number of unordered
choices,
n+r−1
with repetition, of 3 values from 6, so the formula is r
with n =
8


6 and r = 3, that is 3 = 8!/((3!)×(5!)) = (8×7×6)/(3×2×1) =
56.
As noted, the number

ofn+5
outcomes when n indistinguishable dice
n+6−1


are thrown is n
= n .
n+5


Remark By Exercise 11.1.3, this is the same as 5
.

11
 Solutions to Chapter 11 Exercises
in Discrete Mathematics by Norman L. Biggs;
2nd Edition 2002

11.2.3 Suppose that the expression (x + y + z)n is expanded and

the terms are collected, according to the usual rules of algebra: for
example,
(x + y + z)2 = x2 + y 2 + z 2 + 2xy + 2yz + 2zx.
What is the number of terms in the resulting formula?

Discussion Since the terms x, y and z have the properties that

xy = yx, yz = zy and zx = xz (that is, they commute), the
argument in the proof of Theorem 11.2 shows that we are dealing
with unordered selections of n items from 3, with repetition.

Solution The relevant formula is n+r−1 n+2

n
with r = 3, that is n
,
n+2


which by Exercise 11.1.3 is the same as 2
= (n + 1)(n + 2)/2.

12
 Solutions to Chapter 11 Exercises
in Discrete Mathematics by Norman L. Biggs;
2nd Edition 2002

11.3 The Binomial Theorem

11.3.1 Write out the formulae for (1 + x)8 and (1 − x)8 .

Solution We have:

(1 + x)8 = 1 + 8x + 28x2 + 56x3 + 70x4 + 56x5 + 28x6 + 8x7 + x8

and:

(1 − x)8 = 1 − 8x + 28x2 − 56x3 + 70x4 − 56x5 + 28x6 − 8x7 + x8

14
 Solutions to Chapter 11 Exercises
in Discrete Mathematics by Norman L. Biggs;
2nd Edition 2002

11.3.2 Calculate the coefficients of:

(i) x5 in (1 + x)11 ;
(ii) a2 b8 in (a + b)10 ;
(iii) a6 b6 in (a2 + b3 )5 ;
(iv) x3 in (3 + 4x)6 .

Discussion Parts (iii) and (iv) indicate typical difficulties, with

which great care is needed, in obtaining the correct coefficients.
11 11



Solution (i) The coefficient is: 5
= 6

= (11 × 10 × 9 × 8 × 7)/(1 × 2 × 3 × 4 × 5)
= 11 × 3 × 2 × 7 = 462;

10 10



(ii) The coefficient is: 2
= 8
= (10 × 9)/(1 × 2) = 45;
5


(iii) The coefficient is: 3
= (5 × 4 × 3)/(1 × 2 × 3) = 10;

33 × 43 × 63 =


(iv) The coefficient is:

123 × (6 × 5 × 4)/(1 × 2 × 3) = 123 × 20 = 34560.

15
 Solutions to Chapter 11 Exercises
in Discrete Mathematics by Norman L. Biggs;
2nd Edition 2002

11.3.3 Use the identity (1 + x)m (1 + x)n = (1 + x)m+n

to prove that
m+n m n m




n
m n



r
= 0 r
+ 1 r−1
+ . . . + r 0

where m, n, and r are positive integers, m ≥ r, and n ≥ r.

Discussion This is another useful exercise in keeping track of rela-

tions between different binomial numbers.
m+n


Solution The term r
is the coefficient of xm+n in the
m+n
binomial expansion of (1 + x) .
But since m ≥ r and n ≥ r, we can write
the term in
xr in
that expansion as the sum of terms of form ( mi xi ) × ( n−r+i
n
xr−i ),
m
where the first factor is taken from the expansion of (1 + x) , the
second is taken from the expansion of (1 + x)n , and i ranges from
0 to r.
Comparing the two expressions for the term of degree r yields
the formula required.

16
 Solutions to Chapter 11 Exercises
in Discrete Mathematics by Norman L. Biggs;
2nd Edition 2002

11.4 The Sieve Principle

11.4.1 In a class of 67 mathematics students, 47 can read French,
35 can read German and 23 can read both languages. How many
can read neither language? If, furthermore, 20 can read Russian, of
whom 12 also read French, 11 also read German and five read all
three languages, how many cannot read any of the three languages?

Discussion Exercises like this depend on Theorem 11.4 and the

corollary mentioned just after it. When meeting such exercises for
the first time, it is often helpful to fill in the numbers on sketches
resembling those in Fig. 11.1 (Venn diagrams) as well as to approach
the question algebraically.

Solution From the corollary mentioned after Theorem 11.4 we first

wish to calculate the number:

|X \ (AF ∪ AG )| = N − α1 + α2

where AF and AG are the sets of French and of German speakers,

respectively, X of size N is the set of all the mathematics students,
α1 = |AF | + |AG | and α2 = |AF ∩ AG |.

Here N = 67, α1 = 82 and α2 = 23, so N − α1 + α2 = 67 −

82 + 23 = 8.
We also wish to calculate the number:

|X \ (AF ∪ AG ∪ AR | = N − α1 + α2 − α3

where, in addition to the previous notation, AR is the set of Russian

speakers and this time:
α1 = |AF | + |AG | + |AR |,

19
and
α2 = |AF ∩ AG | + |AG ∩ AR | + |AR ∩ AF |,
and
α3 = |AF ∩ AG ∩ AR |.
Here N = 67, α1 = 102, α2 = 23 + 12 + 11 = 46 and α3 = 5,
so N − α1 + α2 − α3 = 67 − 102 + 46 − 5 = 6.

20
 Solutions to Chapter 11 Exercises
in Discrete Mathematics by Norman L. Biggs;
2nd Edition 2002

11.4.2 Find the number of ways of arranging the letters A, E, M,

O, U, Y in a sequence in such a way that the words ME and YOU
do not occur.

Discussion The underlying idea, in an exercise like this, is to cal-

culate the sizes of the sets of all sequences, of all sequences involv-
ing the subsequence ME, of all sequences involving the subsequence
YOU, and of all sequences involving both subsequences. It is then
straightforward to apply Theorem 11.4, and it often helps under-
standing if we start with the smallest of the sets. The idea of a
‘block’ used below is worth keeping in mind in future.

Solution There are six characters, so in a sequence which con-

tains both subsequences, there are two questions to decide: which
of ME and of YOU (reading left-to-right) comes first (two possi-
bilities), and then where the character A goes (three possibilities in
each case). Evidently, therefore, there are 6 sequences which contain
both ‘prohibited’ subsequences.
Alternatively, and more directly, there are three items - A, and
the blocks ME and YOU to arrange in order, hence 3! = 6 possible
sequences.)
To count the sequences which contain the subsequence ME, we
note that their number will simply be the number of sequences of
the four characters A, O, U, Y and the ‘block’ ME, i.e. 5! = 120.
Similarly, to count the sequences which contain the subsequence
YOU, we note that their number will simply be the number of se-
quences of the three characters A, M, E and the ‘block’ YOU, i.e.
4! = 24.
So the number sought is: 6! − (5! + 4!) + 3! = 720 − 144 +
6 = 582.

21
 Solutions to Chapter 11 Exercises
in Discrete Mathematics by Norman L. Biggs;
2nd Edition 2002

11.4.3 Calculate the number d4 of derangements of {1, 2, 3, 4} and

write down the relevant permutations in cycle notation.

Discussion .From p. 114, the number d4 of derangements is the

number of permutations π with the property that π(i) 6= i for
1 ≤ i ≤ 4. The cycle notation is explained on p. 102.

Solution If 0 ≤ j ≤ 4 let Tj be the set of those permutations which

fix exactly j elements of Y = {1, 2, 3, 4}. We wish to calculate |T0 |,
and we note of course that T3 = ∅.
Clearly T4 = {idY } = {(1)(2)(3)(4)}

in cycle notation. If π ∈ S4
4
fixes exactly 2 elements, there are 2 = 6 ways of choosing that
pair, and the effect of π on the other pair is, necessarily, that it
swaps them round, so |T2 | = 6. If π ∈ S4 fixes exactly 1 element,
there are 4 ways of choosing that element, and if, for example, the
fixed element is 1 then π has to be either (1)(234) or (1)(243); it
follows that |T1 | = 8.
Hence |T0 | = 4! − (1 + 6 + 8) = 9.
Evidently, one class of derangements will be those which can be
written as the product of two 2-cycles:
(12)(34), (13)(24), and (14)(23).
If a derangement is not in this class then it must be a 4-cycle, and
the possibilities are:
(1234), (1243), (1324), (1342), (1423) and (1432).
These nine distinct derangements must, therefore, be all of them.

Remark This value (9) can also be found directly from the formula
for dn on p. 114:
dn = n!(1 − 1/(1!) + 1/(2!) − . . . + (−1)n /(n!)).

22
So we have
d4 = 4! × {1 − 1 + (1/2) − (1/6) + (1/24)}
= 24 − 24 + 12 − 4 + 1 = 9.

23
 Solutions to Chapter 11 Exercises
in Discrete Mathematics by Norman L. Biggs;
2nd Edition 2002

11.4.5 Show that the number of derangements of {1, 2, . . . , n} in

which a given object (say 1) is in a 2-cycle is (n − 1)dn−2 . Hence
construct a direct proof of the recursion formula given in the previ-
ous exercise.

Discussion This exercise, of course, assume that n ≥ 3.

Solution Suppose that π is a derangement of {1, 2, . . . , n} and

that π 2 (1) = 1; then the cycle decomposition of π includes a factor
(1 π(1)), that is a 2-cycle. Conversely, if the cycle decomposition
of π includes the 2-cycle (1 m) for some m ∈ {1, 2, . . . , n} then
π 2 (1) = 1.
For simplicity, consider first the case m = 2, so that π = (12)τ
where τ is a derangement of {3, 4, . . . , n}, and there are dn−2 such
derangements, so there are at most dn−2 derangements which include
the ‘factor’ (12). On the other hand, given any derangement τ of
{3, 4, . . . , n}, we obtain a derangement of the type under discussion
by defining π = (12)τ , and clearly distinct derangements τ give
distinct derangements of {1, 2, . . . , n}. It follows that there are
exactly dn−2 derangements which include the factor (12). Since there
are n − 1 possible choices for m in the factor (1m) of our original
derangement of {1, 2, . . . , n} , there are (n − 1)dn−2 derangements
π in which 1 appears in a 2-cycle, or equivalently which satisfy
π 2 (1) = 1.
It remains to count the other derangements of {1, 2, . . . , n}.
Suppose that π is some such deragement, and note that, in the
presentation of π as a product of cycles, 1 appears in a cycle of at
least 3 items, so if we simply delete the 1 from the presentation we
get a derangement of {2, 3, . . . , n}, which we can label π1 . Evi-
dently, we would obtain π1 by this procedure from a derangement
ρ of {1, 2, . . . , n} (not involving 1 in a 2-cycle) if and only if the

25
cycles of ρ1 were the same (up to the ordering of the items in each
cycle) as those of π1 . If we adopt the temporary convention that,
in writing a permutation as a product of cycles, we write each cycle
with the smallest number at the extreme left, and then order the
cycles so that those smallest integers are in increasing order from
left to right, then that presentation is unique. It is therefor clear
that the 1 in the expression for ρ could come before any of the
n − 1 symbols (numbers) in the presentation of π1 , and therefore
that there are precisely n − 1 distinct derangements ρ (not involving
1 in a 2-cycle) of {1, 2, . . . , n} which yield the same derangement
of {2, 3, . . . , n}.
The same argument about ‘insertion’ of a 1 before any symbol
of a derangement of {2, 3, . . . , n}, show that every such derange-
ment can be obtained as π1 for a suitable choice of π, so that the
function π −→ π1 is a surjection. There are therefore (n − 1)dn−1
derangements of {1, 2, . . . , n} which do involve 1 in a 2-cycle, and
the required formula then follows at once.

26
 Solutions to Chapter 11 Exercises
in Discrete Mathematics by Norman L. Biggs;
2nd Edition 2002

11.5 Some Arithmetical Applications

11.5.1 Calculate φ(n) and µ(n) for each n in the range 95 ≤ n ≤
100.

Discussion The relevant definitions are on p. 95 and p. 117 respec-

tively. In order to calculate either, for a specific choice of n ∈ N, it
is necessary to find the prime factorization of n, and then the result
of Exercise 3 (below) can routinely be used to simplify calculations.
Note also that if p, e ∈ N and p is prime then, from Theorem 11.5.1,
φ(pe ) = pe−1 (p − 1).

Solution We know that 95 = 5 × 19, so φ(95) = φ(5) × φ(19) =

4 × 18 = 72, and µ(95) = (−1)2 = 1 .
Similarly, 96 = 25 × 3, so φ(96) = φ(32) × φ(3) = 16 × 2 = 32,
and µ(96) = 0.
Since 97 is prime, φ(97) = 96 and µ(97) = −1.
We know that 98 = 2 × 72 , so φ(98) = φ(2) × φ(72 ) = 1 × 42 =
42, and µ(98) = 0.
Similarly, 99 = 32 × 11, so φ(99) = φ(9) × φ(11) = 6 × 10 = 60,
and µ(99) = 0.
Finally, 100 = 22 ×52 , so φ(100) = φ(4)×φ(25) = 2×20 = 40,
and µ(100) = 0.

27
 Solutions to Chapter 11 Exercises
in Discrete Mathematics by Norman L. Biggs;
2nd Edition 2002

11.5.2 Show that if the prime factorization of n is pe11 pe22 . . . perr then
the number of divisors of n is
(e1 + 1)(e2 + 1) . . . (er + 1)
Solution From Theorem 8.6.1 (unique prime factorization), d ∈ N
is a divisor of n if and only if each prime factor (if any) of d is a
prime factor of n, and no such prime, for example pj , is a factor of
d to a higher power than its power ej as a factor of n.
That is, the factors d of n are precisely the numbers
pf11 pf22 . . . pfrr with 0 ≤ fj ≤ ej for 1 ≤ j ≤ r,
the number of which is clearly (e1 + 1)(e2 + 1) . . . (er + 1)

Remark The trivial divisors of n are 1 and n itself, which correspond

respectively to the cases
f1 = f2 = . . . = fr = 0 and f1 = e1 ; f2 = e2 ; . . . ; fr = er .
Thus n is prime if and only if n > 1 and its only divisors are the
trivial ones.

28
 Solutions to Chapter 11 Exercises
in Discrete Mathematics by Norman L. Biggs;
2nd Edition 2002

11.5.3 Use Theorem 11.5.1 to show that if gcd (m, n) = 1 then

φ(mn) = φ(m)φ(n).

Discussion From Theorem 8.6.1 (unique prime factorization), it is

clear that, for m, n ∈ N, gcd (m, n) = 1 if and only if m and n have
no common prime factor.

Solution Suppose that p1 , . . . , pr , q1 , . . . qs are primes and

e1 , . . . er , f1 , . . . , fs
are non-negative integers (which will all be positive except if m = 1
and/or n = 1), and that
m = pe11 pe22 . . . perr and n = q1f1 q2f2 . . . qrfr .

Since m and n have no prime factor in common, by Theorem 11.5.1:

φ(m) = m × (1 − p−1 −1 −1
1 ) × (1 − p2 ) × . . . × (1 − pr ), and
φ(n) = n × (1 − q1−1 ) × (1 − q2−1 × . . . × (1 − qr−1 ), and
φ(mn) = m × (1 − p−1 −1 −1
1 ) × (1 − p2 ) × . . . × (1 − pr )×
−1 −1 −1
n × (1 − q1 ) × (1 − q2 × . . . × (1 − qr ).
The last expression is easily seen to be equal to that given by
Theorem 11.5.1 for φ(mn).

Remark It is also possible to prove the result directly from the

definitions.

29
 Solutions to Chapter 11 Exercises
in Discrete Mathematics by Norman L. Biggs;
2nd Edition 2002

11.6 Designs
11.6.1 Undeterred by the apparent failure of his scheme to force
every student to take exactly four of the seven mathematics courses
(Section 10.2), Professor McBrain is working on a plan to ensure
that, in addition, each course has the same number of takers.
(i) If there are v students, and k take each course, what is the
relationship between v and k?
(ii) If the actual number of students is 53, how many must
Professor McBrain expel in order to make his plan feasible?
(iii) Write down explicit designs for McBrain’s plan in the cases
v = 7 and v = 14. (Remember that in a design no block can be
repeated - that is, no two courses can have the same set of takers.)

Discussion Practising devising designs with a given triple of param-

eters is a good way to convinve oneself that Theorem 11.6 is indeed
a remarkable result.

Solution (i) Arguing as in the discussion of Table 11.6.1, we find

that 4v = 7k.
(ii) If e is the number of students to be expelled, then the pro-
fessor will have v = 53 − e students left, so that 4 × (53 − e), and
therefore 53 − e, must be divisible by 7. Thus e must be at least 4.
(iii) If v = 7 then k = 4 (from (i)), so one possible solution is:
{4, 5, 6, 7}, {1, 5, 6, 7}, {1, 2, 6, 7}, {1, 2, 3, 7},
{1, 2, 3, 4}, {2, 3, 4, 5}, {3, 4, 5, 6}.

If v = 14 then k = 8. One possible solution is:

{7, 8, 9, 10, 11, 12, 13, 14}, {2, 3, 4, 5, 6, 12, 13, 14},

34
{3, 4, 5, 6, 8, 9, 10, 11}, {1, 3, 5, 6, 8, 10, 12, 14},
{1, 2, 5, 6, 7, 9, 12, 13}, {1, 2, 4, 7, 9, 11, 13, 14},
{1, 2, 3, 4, 7, 8, 10, 11}.

35
 Solutions to Chapter 11 Exercises
in Discrete Mathematics by Norman L. Biggs;
2nd Edition 2002

11.6.2 For the following values of (v, k, r) either construct a design

with those parameters or explain why such a design cannot exist.
(i) (v, k, r) = (6, 3, 1); (ii) (v, k, r) = (5, 2, 1);
(iii) (v, k, r) = (7, 3, 3); (iv) (v, k, r) = (9, 6, 4).

Solution (i) Since b = 6 × 1/3 = 2, it is easy to see that

{1, 2, 3}, {4, 5, 6} will do.
(ii) Since 2 - (5 × 1), there cannot be a design with parameters
(5, 2, 1).
(iii) Since 3 | (7 × 3) and vr/k = 7 < 73 = 35, there must be


a design fitting these parameters. One possibility is:
{1, 2, 3}, {2, 3, 4}, {3, 4, 5}, {4, 5, 6},
{5, 6, 7}, {1, 6, 7}, {1, 2, 7}.
(iv) Since 6 | (9 × 4) and vr/k = 6 < 96 = 84, there must be


a design fitting these parameters. One possibility is:
{1, 2, 3, 4, 5, 6}, {1, 2, 3, 7, 8, 9}, {4, 5, 6, 7, 8, 9},
{1, 2, 4, 5, 7, 8}, {1, 3, 4, 6, 7, 9}, {2, 3, 5, 6, 8, 9}.

36
 Solutions to Chapter 11 Exercises
in Discrete Mathematics by Norman L. Biggs;
2nd Edition 2002

11.6.3 What is the value of r in the design whose blocks are all
the k-subsets of a v-set?
v


Solution There are k
k-subsets of a v-set, so r must satisfy the
equation
v


k ×k = v × r, i.e.,
(v!) × k = v × r × (k!) × ((v − k)!), i.e. ,
((v − 1)!) = r × ((k − 1)!) × (((v − 1) − (k − 1))!),

v−1


so that r = k−1 .

37
 Solutions to Chapter 11 Exercises
in Discrete Mathematics by Norman L. Biggs;
2nd Edition 2002

11.7 t-designs

11.7.1 Given that there is a 5-design with parameters v = 12,

k = 6, and r5 = 1, find the values of r4 , r3 , r2 , r1 and b.

Discussion The relevant formulae are given by Theorem 11.7.2, in

particular the special case:
rt−1 = rt × {(v − t + 1)/(k − t + 1)} .
Solution From the formula just given,
r4 = 1 × {(12 − 5 + 1)/(6 − 5 + 1)} = 8/2 = 4 ;
r3 = 4 × {(12 − 4 + 1)/(6 − 4 + 1)} = 36/3 = 12 ;
r2 = 12 × {(12 − 3 + 1)/(6 − 3 + 1)} = 120/4 = 30 ;
r1 = 30 × {(12 − 2 + 1)/(6 − 2 + 1)} = 330/5 = 66 ;

and b = 66 × {(12 − 1 + 1)/(6 − 1 + 1)} = 66 × 2 = 132 .

39
 Solutions to Chapter 11 Exercises
in Discrete Mathematics by Norman L. Biggs;
2nd Edition 2002

11.7.2 Is it possible that designs of the following kind exist?

(i) A 3-design with v = 15, k = 6, and r3 = 2.
(ii) A 4-design with v = 11, k = 5, and r4 = 1.

Discussion This exercise involves checking whether the given pa-

rameters are capable of satisfying the equations which follow from
Theorem 11.7.2.

Solution (i) If there were such a design, then there would be a

2-design with
r2 = r3 × {(15 − 3 + 1)/(6 − 3 + 1)} = 2 × (13/4) = 6.5 ;
which is impossible since r2 has to be an integer (in fact, a positive
natural number).
(ii) If there is such a design, then there are: a 3-design, a 2-design
and a 1-design with
r3 = r4 × {(11 − 4 + 1)/(5 − 4 + 1)} = 1 × (8/2) = 4 ;
r2 = r3 × {(11 − 3 + 1)/(5 − 3 + 1)} = 4 × (9/3) = 12 ;
r1 = r2 × {(11 − 2 + 1)/(5 − 2 + 1)} = 12 × (10/4) = 30 .
The existence of such a design would also mean that
b = r0 = r1 × {(11 − 1 + 1)/(5 − 1 + 1)} = 30 × (11/5) = 66 .

Looking at Ex. 1, the 5-design mentioned there yields a 4-design

with the required properties.

40
 Solutions to Chapter 11 Exercises
in Discrete Mathematics by Norman L. Biggs;
2nd Edition 2002

11.7.3 A 2-design with k = 3 and r2 = 1 is usually called a

Steiner triple system (STS). Determine for which values of v in the
range 3 ≤ v ≤ 12 an STS with v varieties can exist, and construct
an STS in these cases.

Discussion This is another exercise which involves checking whether

the possible parameters are capable of satisfying the equations which
follow from Theorem 11.7.2, plus some detailed work to find satis-
factory examples.

Solution If an STS exists for v then there exist r1 and r0 = b such

that
r1 = r2 × {(v − 1)/(k − 1)} = (v − 1)/2 ∈ N ; (1)
b = r0 = r1 × v/k ∈ N. (2)
From (1), v − 1 must be even, i.e. v must be odd. We can suppose
that v = 6n − i where n ∈ N, i ∈ {1, 3, 5} and (for the given range
of values of v) n = 1 or n = 2. From (2), r0 = r1 × (v/k) =
v(v − 1)/6 , so that {(6n − i) × (6n − i − 1)/6} ∈ N, and hence
(i(i + 1)/6) ∈ N. It follows that i = 3 or i = 5, and therefore (for
3 ≤ v ≤ 12) that v = 3 ( with n = 1) or v = 7 or v = 9 ( with
n = 2).
The following 2-designs are suitable STS’s for these three values
of v
v = 3: {1, 2, 3};

v = 7: {1, 2, 4}, {2, 3, 5}, {3, 4, 6}, {4, 5, 7},

{5, 6, 1}, {6, 7, 2}, {7, 1, 3};

v = 9: {1, 2, 3}, {1, 4, 7}, {1, 5, 9}}, {1, 6, 8},

{4, 5, 6}, {2, 5, 8}, {2, 6, 7}, {2, 4, 9},
{7, 8, 9}, {3, 6, 9}, {3, 4, 8}, {3, 5, 7}.

41