Skip to main content

Section Solutions for Chapter 18

Solutions to Exercise 18.1.12

18.1.12.2 Since \(\displaystyle v = 39 = 6 \cdot 6 + 3 \equiv 3 \pmod{6}\text{,}\) the proof of Theorem 18.1.8 tells us that we should use a Latin square constructed in Lemma 18.1.4. Since \(\displaystyle v/3 = 39/3 = 13\text{,}\) the Latin square is of order \(\displaystyle n = 13\text{,}\) so the first sentence of the proof of Lemma 18.1.4 tells us that the first row of the Latin square is
\begin{equation*} \begin{matrix} 1 \amp \displaystyle \frac{13 + 3}{2} \amp 2 \amp \displaystyle \frac{13 + 5}{2} \amp 3 \amp \displaystyle \frac{13 + 7}{2} \amp 4 \amp \cdots \amp 13 \amp \displaystyle\frac{13 + 1}{2} \end{matrix}\text{.} \end{equation*}
In other words, the first row is
\begin{equation*} \begin{matrix} 1 \amp 8 \amp 2 \amp 9 \amp 3 \amp 10 \amp 4 \amp 11 \amp 5 \amp 12 \amp 6 \amp 13 \amp 7 \end{matrix}\text{.} \end{equation*}
Then the second sentence of the proof of Lemma 18.1.4 tells us that the rest of the rows are obtained by shifting to the left. So the Latin square is
\begin{equation*} \begin{array}{rrrrrrrrrrrrr} 1 \amp 8 \amp 2 \amp 9 \amp 3 \amp 10 \amp 4 \amp 11 \amp 5 \amp 12 \amp 6 \amp 13 \amp 7 \\ 8 \amp 2 \amp 9 \amp 3 \amp 10 \amp 4 \amp 11 \amp 5 \amp 12 \amp 6 \amp 13 \amp 7 \amp 1 \\ 2 \amp 9 \amp 3 \amp 10 \amp 4 \amp 11 \amp 5 \amp 12 \amp 6 \amp 13 \amp 7 \amp 1 \amp 8 \\ 9 \amp 3 \amp 10 \amp 4 \amp 11 \amp 5 \amp 12 \amp 6 \amp 13 \amp 7 \amp 1 \amp 8 \amp 2 \\ 3 \amp 10 \amp 4 \amp 11 \amp 5 \amp 12 \amp 6 \amp 13 \amp 7 \amp 1 \amp 8 \amp 2 \amp 9 \\ 10 \amp 4 \amp 11 \amp 5 \amp 12 \amp 6 \amp 13 \amp 7 \amp 1 \amp 8 \amp 2 \amp 9 \amp 3 \\ 4 \amp 11 \amp 5 \amp 12 \amp 6 \amp 13 \amp 7 \amp 1 \amp 8 \amp 2 \amp 9 \amp 3 \amp 10 \\ 11 \amp 5 \amp 12 \amp 6 \amp 13 \amp 7 \amp 1 \amp 8 \amp 2 \amp 9 \amp 3 \amp 10 \amp 4 \\ 5 \amp 12 \amp 6 \amp 13 \amp 7 \amp 1 \amp 8 \amp 2 \amp 9 \amp 3 \amp 10 \amp 4 \amp 11 \\ 12 \amp 6 \amp 13 \amp 7 \amp 1 \amp 8 \amp 2 \amp 9 \amp 3 \amp 10 \amp 4 \amp 11 \amp 5 \\ 6 \amp 13 \amp 7 \amp 1 \amp 8 \amp 2 \amp 9 \amp 3 \amp 10 \amp 4 \amp 11 \amp 5 \amp 12 \\ 13 \amp 7 \amp 1 \amp 8 \amp 2 \amp 9 \amp 3 \amp 10 \amp 4 \amp 11 \amp 5 \amp 12 \amp 6 \\ 7 \amp 1 \amp 8 \amp 2 \amp 9 \amp 3 \amp 10 \amp 4 \amp 11 \amp 5 \amp 12 \amp 6 \amp 13 \end{array} \end{equation*}

18.1.12.4 No, it is not a Kirkman system.

We will call \(\displaystyle \{u_1, \ldots,u_5\}\) the \(\displaystyle u\)-girls; \(\displaystyle \{v_1, \ldots, v_5\}\) the \(\displaystyle v\)-girls; and \(\displaystyle \{w_1, \ldots, w_5\}\) the \(\displaystyle w\)-girls. Of the \(\displaystyle 35\) blocks that we obtained through the construction, \(\displaystyle 5\) have one \(\displaystyle u\)-girl, one \(\displaystyle v\)-girl, and one \(\displaystyle w\)-girl; the other \(\displaystyle 30\) have either two \(\displaystyle u\)-girls with a \(\displaystyle v\)-girl, two \(\displaystyle v\)-girls with a \(\displaystyle w\)-girl, or two \(\displaystyle w\)-girls with a \(\displaystyle u\)-girl.

A Kirkman system requires us to divide the blocks into \(\displaystyle 7\) groups of \(\displaystyle 5\) blocks such that each girl appears exactly once in each group of blocks. Since there should be \(\displaystyle 7\) groups of \(\displaystyle 5\) blocks, but there are only \(\displaystyle 5\) blocks that have a \(\displaystyle u\)-girl, a \(\displaystyle v\)-girl, and a \(\displaystyle w\)-girl, there must be at least one group of blocks (in fact, at least two) that has no block consisting of a \(\displaystyle u\)-girl, a \(\displaystyle v\)-girl, and a \(\displaystyle w\)-girl.

Consider such a group of \(\displaystyle 5\) blocks. We must have all \(\displaystyle 5\) of the \(\displaystyle u\)-girls. If no block contained more than one \(\displaystyle u\)-girl, then in order to get all \(\displaystyle 5\) \(\displaystyle u\)-girls we would have to choose only blocks that have two \(\displaystyle w\)-girls and a \(\displaystyle u\)-girl. However, this would mean that we had \(\displaystyle 10\) \(\displaystyle w\)-girls and no \(\displaystyle v\)-girls, which is not allowed. So we must choose at least one block that has two \(\displaystyle u\)-girls and a \(\displaystyle v\)-girl. Repeating the same argument with \(\displaystyle v\) or \(\displaystyle w\) taking the place of \(\displaystyle u\text{,}\) we see that we must also choose at least one block that has two \(\displaystyle v\)-girls and a \(\displaystyle w\)-girl, and at least one block that has two \(\displaystyle w\)-girls and a \(\displaystyle u\)-girl. Since we are only choosing \(\displaystyle 5\) blocks but there are these three classes of blocks, there must be some class of blocks of which we only choose one.

Without loss of generality, suppose that we only choose one of the blocks that has two \(\displaystyle u\)-girls and a \(\displaystyle w\)-girl. In order to have all \(\displaystyle 5\) of the \(\displaystyle u\)-girls, we must choose three blocks that have two \(\displaystyle w\)-girls and a \(\displaystyle u\)-girl. But this means that we have six \(\displaystyle w\)-girls, which is not allowed.

Therefore, there is no way to partition the blocks of this design into seven groups of five blocks so that every girl appears exactly once in each group.

Solutions to Exercise 18.2.7

18.2.7.2 We must have \(\displaystyle b\binom{k}{t}=\lambda\binom{v}{t}=\binom{15}{t}\text{.}\)

Since we are not including any trivial \(\displaystyle t-(v,t,1)\) design, we have \(\displaystyle t \ge 2\text{,}\) \(\displaystyle 3 \le k \le 14\text{,}\) and \(\displaystyle t\lt k\text{.}\)

Now

\begin{equation*} \frac{15!}{t!(15-t)!}=b\frac{k!}{t!(k-t)!}\text{,} \end{equation*}

which means that

\begin{equation*} \frac{15\cdot 14 \cdots (16-t)}{k(k-1)\cdots (k+1-t)} \end{equation*}

is an integer.

Furthermore, we have \(\displaystyle \binom{k-1}{t-1}\) divides \(\displaystyle \binom{14}{t-1}\text{,}\) so that \(\displaystyle \frac{(k-1)!}{(k-t)!}\) divides \(\displaystyle \frac{14!}{(15-t)!}\text{.}\) In other words,

\begin{equation*} \frac{14!(k-t)!}{(15-t)!(k-1)!}=\frac{14\cdot 13 \cdots (16-t)}{(k-1)(k-2)\cdots(k+1-t)} \end{equation*}

is an integer. If we call this integer \(\displaystyle y\text{,}\) combining this with the previous paragraph tells us that \(\displaystyle k\) is a divisor of \(\displaystyle 15y\text{.}\) We can also further work with the algebra to obtain

\begin{equation*} y=\frac{14\cdot 13\cdots k}{(15-t)(14-t)\cdots(k+1-t)}\text{.} \end{equation*}

When \(\displaystyle k=14\text{,}\) this gives \(\displaystyle y=14/(15-t)\text{.}\) Since \(\displaystyle k\) divides \(\displaystyle 15y\) and \(\displaystyle k\) is coprime to \(\displaystyle 15\text{,}\) we must have \(\displaystyle k\) divides \(\displaystyle y\text{.}\) But then \(\displaystyle y/14=1/(15-t)\) is an integer, implying \(\displaystyle t=14\text{.}\) This contradicts \(\displaystyle t\lt k\text{.}\) Thus \(\displaystyle k=14\) cannot arise.

When \(\displaystyle k=13\) this gives \(\displaystyle y=\frac{14 \cdot 13}{(15-t)(14-t)}\text{.}\) Since \(\displaystyle k\) divides \(\displaystyle 15y\) and \(\displaystyle k\) is coprime to \(\displaystyle 15\text{,}\) we must have \(\displaystyle k\) divides \(\displaystyle y\text{.}\) But then \(\displaystyle \frac{y}{13}=\frac{14}{(15-t)(14-t)}\) is an integer. Since \(\displaystyle t\lt k=13\text{,}\) we have \(\displaystyle 14-t \ge 2\text{,}\) but no two consecutive integers each of which is at least \(\displaystyle 2\) are both divisors of \(\displaystyle 14\text{,}\) a contradiction. Thus \(\displaystyle k=13\) cannot arise.

When \(\displaystyle k=12\text{,}\) this gives

\begin{equation*} y=\frac{14 \cdot 13 \cdot 12}{(15-t)(14-t)(13-t)}\text{.} \end{equation*}

Now \(\displaystyle k\) dividing \(\displaystyle 15y\) implies that

\begin{equation*} \frac{15 \cdot 14 \cdot 13}{(15-t)(14-t)(13-t)} \end{equation*}

is an integer. Since the numerator is not a multiple of \(\displaystyle 2^2\text{,}\) the denominator cannot be either, leaving only the possibilities \(\displaystyle t=4,8\text{.}\) Since the numerator is not a multiple of \(\displaystyle 3^2\text{,}\) the denominator cannot be either, which eliminates \(\displaystyle t=4\text{.}\) When \(\displaystyle t=8\text{,}\) the numerator of \(\displaystyle y\) is not a multiple of \(\displaystyle 5\text{,}\) but the denominator is, so this is also impossible. Thus \(\displaystyle k=12\) cannot arise.

When \(\displaystyle k=11\) this gives

\begin{equation*} y=\frac{14\cdot 13 \cdot 12 \cdot 11}{(15-t)(14-t)(13-t)(12-t)}\text{.} \end{equation*}

Since \(\displaystyle k\) divides \(\displaystyle 15y\) and \(\displaystyle k\) is coprime to \(\displaystyle 15\text{,}\) we must have \(\displaystyle k\) divides \(\displaystyle y\text{.}\) But then

\begin{equation*} \frac{y}{11}= \frac{14\cdot 13 \cdot 12}{(15-t)(14-t)(13-t)(12-t)} \end{equation*}

is an integer. Since the numerator is not a multiple of \(\displaystyle 5\text{,}\) the four consecutive numbers that are the factors of the denominator must be \(\displaystyle 6\) through \(\displaystyle 9\) (since \(\displaystyle t \ge 2\text{,}\) they cannot be \(\displaystyle 11\) through \(\displaystyle 14\text{,}\) and since \(\displaystyle t\lt 11\) they cannot be \(\displaystyle 1\) through \(\displaystyle 4\)). Thus, we must have \(\displaystyle t=6\text{.}\) But then the numerator is not divisible by \(\displaystyle 3^2\text{,}\) while the denominator is divisible by \(\displaystyle 3^3\text{,}\) contradicting \(\displaystyle y/11\) being an integer. Thus \(\displaystyle k=11\) is not possible.

When \(\displaystyle k=10\text{,}\) this gives

\begin{equation*} y=\frac{14 \cdot 13 \cdot 12 \cdot 11 \cdot 10}{(15-t)(14-t)(13-t)(12-t)(11-t)}\text{.} \end{equation*}

Now \(\displaystyle k\) dividing \(\displaystyle 15y\) implies that

\begin{equation*} \frac{15 \cdot 14 \cdot 13\cdot 12 \cdot 11}{(15-t)(14-t)(13-t)(12-t)(11-t)} \end{equation*}

is an integer. Since the numerator is not a multiple of \(\displaystyle 2^4\text{,}\) the denominator cannot be either. In particular, \(\displaystyle 8\) cannot be one of the factors that appears in the denominator (since some other even factor would appear with it), nor can \(\displaystyle 2\text{,}\) \(\displaystyle 4\text{,}\) and \(\displaystyle 6\) all be factors that appear in the denominator. Also, the numerator is not divisible by \(\displaystyle 3^3\text{,}\) so we cannot have \(\displaystyle 11-t=9\text{.}\) This leaves \(\displaystyle t=8\) as the only possibility. However, the numerator of \(\displaystyle y\) is not divisible by \(\displaystyle 3^2\text{,}\) so \(\displaystyle t=8\) is also not possible. Thus \(\displaystyle k=10\) is not possible.

When \(\displaystyle k=9\text{,}\) we see that

\begin{equation*} y=\frac{14 \cdot 13 \cdot 12 \cdot 11 \cdot 10\cdot 9}{(15-t)(14-t)(13-t)(12-t)(11-t)(10-t)}\text{.} \end{equation*}

So \(\displaystyle k\) dividing \(\displaystyle 15y\) gives

\begin{equation*} \frac{15\cdot14\cdot13\cdot12\cdot11\cdot10}{(15-t)(14-t)(13-t)(12-t)(11-t)(10-t)} \end{equation*}

being an integer. Since the numerator is not divisible by \(\displaystyle 2^5\text{,}\) the denominator cannot be either. In particular, \(\displaystyle 8\) cannot appear as one of the factors in the denominator (or two other numbers divisible by \(\displaystyle 2\) would also appear as factors), so the only possibility is \(\displaystyle t=8\text{.}\) However, if we take \(\displaystyle k=9\text{,}\) \(\displaystyle t=8\text{,}\) and \(\displaystyle i=2\text{,}\) the necessary condition is \(\displaystyle \binom{7}{6}=7\) divides \(\displaystyle \binom{13}{6}\text{,}\) which is not true. Thus, \(\displaystyle k=9\) is not possible.

When \(\displaystyle k=8\text{,}\) we calculate

\begin{equation*} y=\frac{14\cdot13\cdot12\cdot11\cdot10\cdot9\cdot8}{(15-t)(14-t)(13-t)(12-t)(11-t)(10-t)(9-t)}\text{.} \end{equation*}

Since \(\displaystyle k\) divides \(\displaystyle 15y\) and \(\displaystyle k\) is coprime to \(\displaystyle 15\text{,}\) we must have \(\displaystyle k\) divides \(\displaystyle y\text{.}\) But then

\begin{equation*} \frac{y}{8}=\frac{14\cdot13\cdot12\cdot11\cdot10\cdot9}{(15-t)(14-t)(13-t)(12-t)(11-t)(10-t)(9-t)}\text{.} \end{equation*}

Since the consecutive factors in the denominator include \(\displaystyle 8\) and at least two other even numbers, this implies that the numerator should also be a multiple of \(\displaystyle 2^5\text{,}\) but it is not. Thus \(\displaystyle k=8\) is not possible.

When \(\displaystyle k=7\) we see that

\begin{equation*} y=\frac{14\cdot13\cdot12\cdot11\cdot10\cdot9\cdot8\cdot 7}{(15-t)(14-t)(13-t)(12-t)(11-t)(10-t)(9-t)(8-t)}\text{.} \end{equation*}

Since the consecutive factors in the denominator include \(\displaystyle 6\) and \(\displaystyle 9\text{,}\) if they also include another multiple of \(\displaystyle 3\) then the numerator must be divisible by \(\displaystyle 3^4\text{,}\) but it is not. This leaves the possibility that \(\displaystyle t=4\) so the factors in the denominator are \(\displaystyle 4\) through \(\displaystyle 11\text{,}\) but this is divisible by \(\displaystyle 5^2\text{,}\) which the numerator is not. Thus, \(\displaystyle k=7\) is not possible.

If \(\displaystyle k=6\) then

\begin{equation*} y=\frac{14\cdot13\cdot12\cdot11\cdot10\cdot9\cdot8\cdot 7\cdot 6}{(15-t)(14-t)(13-t)(12-t)(11-t)(10-t)(9-t)(8-t)(7-t)}\text{.} \end{equation*}

So \(\displaystyle k\) dividing \(\displaystyle 15y\) gives

\begin{equation*} \frac{15\cdot14\cdot13\cdot12\cdot11\cdot10\cdot9\cdot8\cdot7}{(15-t)(14-t)(13-t)(12-t)(11-t)(10-t)(9-t)(8-t)(7-t)} \end{equation*}

being an integer. The numerator is not divisible by \(\displaystyle 2^8\text{,}\) so the denominator cannot be either; in particular it cannot include as factors all of the even integers from \(\displaystyle 4\) through \(\displaystyle 10\) as well as one other. This leaves the possibilities \(\displaystyle t=2\) and \(\displaystyle t=4\text{.}\) If \(\displaystyle t=2\) then \(\displaystyle y=14/5\) which is not an integer, and similarly if \(\displaystyle t=4\) then we have \(\displaystyle y=\frac{14\cdot 13\cdot 12}{5\cdot 4\cdot 3}\) which is not an integer. So \(\displaystyle k=6\) is not possible.

If \(\displaystyle k=5\) then

\begin{equation*} y=\frac{14\cdot13\cdot12\cdot11\cdot10\cdot9\cdot8\cdot7\cdot6\cdot 5}{(15-t)(14-t)(13-t)(12-t)(11-t)(10-t)(9-t)(8-t)(7-t)(6-t)}\text{.} \end{equation*}

The numerator is not divisible by \(\displaystyle 2^9\text{,}\) so the denominator cannot be either; in particular, the denominator cannot include all of \(\displaystyle 4\text{,}\) \(\displaystyle 6\text{,}\) \(\displaystyle 8\text{,}\) \(\displaystyle 10\text{,}\) and \(\displaystyle 12\) as factors in the product. This leaves only the possibility \(\displaystyle t=4\text{.}\) If \(\displaystyle k=5\) and \(\displaystyle t=4\) then \(\displaystyle i=0\) gives \(\displaystyle \binom{5}{4}=5\) divides \(\displaystyle \binom{15}{4}=1365\) which is true; \(\displaystyle i=1\) gives \(\displaystyle \binom{4}{3}=4\) divides \(\displaystyle \binom{14}{3}=364\) which is true; \(\displaystyle i=2\) gives \(\displaystyle \binom{3}{2}=3\) divides \(\displaystyle \binom{13}{2}=78\text{,}\) which is true; \(\displaystyle i=3\) gives \(\displaystyle \binom{2}{1}=2\) divides \(\displaystyle \binom{12}{1}=12\text{,}\) which is true. Thus a \(\displaystyle 4-(15,5,1)\) design could exist, but this is the only possibility with \(\displaystyle k=5\text{.}\)

When \(\displaystyle k=4\) we have

\begin{equation*} y=\frac{14\cdot13\cdot12\cdot11\cdot10\cdot9\cdot8\cdot7\cdot6\cdot5\cdot 4}{(15-t)(14-t)(13-t)(12-t)(11-t)(10-t)(9-t)(8-t)(7-t)(6-t)(5-t)}\text{.} \end{equation*}

The denominator includes \(\displaystyle 3\text{,}\) \(\displaystyle 6\text{,}\) \(\displaystyle 9\text{,}\) and \(\displaystyle 12\text{,}\) so is divisible by \(\displaystyle 3^5\text{,}\) but the numerator is not. Thus, \(\displaystyle k=4\) is not possible.

If \(\displaystyle k=3\) then \(\displaystyle 2 \le t\lt k\) implies \(\displaystyle t=2\text{.}\) We know these parameters are possible, as these are the parameters of a Steiner triple system.

Thus, the only possible values of \(\displaystyle k\) and \(\displaystyle t \ge 2\) for which nontrivial \(\displaystyle t\)-designs might exist with \(\displaystyle v=15\) and \(\displaystyle \lambda=1\) are \(\displaystyle k=5\) and \(\displaystyle t=4\text{:}\) a \(\displaystyle 4-(15,5,1)\) design, or \(\displaystyle k=3\) and \(\displaystyle t=2\text{:}\) a \(\displaystyle (15,3,1)\) Steiner triple system.

18.2.7.3 If a \(\displaystyle 3-(16,6,1)\) design exists then we have \(\displaystyle \binom{6-i}{3-i}\) divides \(\displaystyle \binom{16-i}{3-i}\) for \(\displaystyle 0 \le i \le 2\text{.}\) With \(\displaystyle i=0\) this gives \(\displaystyle \binom{6}{3}=20\) divides \(\displaystyle \binom{16}{3}=560\text{,}\) which is true. With \(\displaystyle i=1\text{,}\) we have \(\displaystyle \binom{5}{2}=10\) divides \(\displaystyle \binom{15}{2}=105\text{,}\) which is not true. Therefore such a design is not possible.

Solutions to Exercise 18.3.9

18.3.9.1 Proof. Let \(\displaystyle L\text{,}\) \(\displaystyle M\text{,}\) and \(\displaystyle N\) be lines of an affine plane such that \(\displaystyle L\) is parallel to both \(\displaystyle M\) and \(\displaystyle N\text{;}\) that is, no point lies on both \(\displaystyle L\) and \(\displaystyle M\) or on both \(\displaystyle L\) and \(\displaystyle N\text{.}\) Let \(\displaystyle p\) be an arbitrary point on \(\displaystyle M\text{.}\) Since \(\displaystyle L\) and \(\displaystyle M\) are parallel, \(\displaystyle p\) does not lie on \(\displaystyle L\text{.}\) By the parallel postulate, there is a unique line through \(\displaystyle p\) that is parallel to \(\displaystyle L\text{;}\) this line is \(\displaystyle M\text{.}\) Therefore \(\displaystyle N\) cannot contain \(\displaystyle p\text{.}\) Since our choice of \(\displaystyle p\) on \(\displaystyle M\) was arbitrary, no point of \(\displaystyle M\) can also lie on \(\displaystyle N\text{,}\) so \(\displaystyle M\) and \(\displaystyle N\) are parallel. ◾

18.3.9.3 A finite affine plane of order \(\displaystyle 19\) has \(\displaystyle 19^2=361\) points, and \(\displaystyle 19(19+1)=380\) lines.

18.3.9.5 Without colours it is difficult to effectively draw this plane so that the parallel classes can be clearly seen. We will use solid vertical lines; solid horizontal lines; dashed lines; dotted lines; solid grey lines; and dotted grey lines to represent the six parallel classes of lines, but some of these may be difficult to distinguish. Note that the lines that are neither vertical nor horizontal will “turn corners” or zig-zag to join their sets of \(\displaystyle 5\) points.
We obtain the following \(\displaystyle 4\) MOLS of order \(\displaystyle 5\) from this affine plane, by using the vertical and horizontal lines to create the coordinates. To make things easier to see, we will have the positions in the Latin squares correspond visually to the positions in the \(\displaystyle 5\) by \(\displaystyle 5\) array of points that we have drawn, so the top-left entry in the Latin squares will come from the top-left point of the array, etc. We will number the lines in each parallel class so as to ensure that the entries in the top row of each square are \(\displaystyle 1\text{,}\) \(\displaystyle 2\text{,}\) \(\displaystyle 3\text{,}\) \(\displaystyle 4\text{,}\) and \(\displaystyle 5\text{,}\) in that order. The first square corresponds to the dashed lines; the second to the dotted lines; the third to the solid grey lines, and the fourth to the dashed grey lines.
\begin{equation*} \begin{matrix}\begin{matrix}1\amp 2\amp 3\amp 4\amp 5\\ 2\amp 3\amp 4\amp 5\amp 1\\ 3\amp 4\amp 5\amp 1\amp 2\\ 4\amp 5\amp 1\amp 2\amp 3\\ 5\amp 1\amp 2\amp 3\amp 4 \end{matrix} \quad \amp \begin{matrix}1\amp 2\amp 3\amp 4\amp 5\\ 5\amp 1\amp 2\amp 3\amp 4\\ 4\amp 5\amp 1\amp 2\amp 3\\ 3\amp 4\amp 5\amp 1\amp 2\\ 2\amp 3\amp 4\amp 5\amp 1 \end{matrix} \quad \amp \begin{matrix}1\amp 2\amp 3\amp 4\amp 5\\ 3\amp 4\amp 5\amp 1\amp 2\\ 5\amp 1\amp 2\amp 3\amp 4\\ 2\amp 3\amp 4\amp 5\amp 1\\ 4\amp 5\amp 1\amp 2\amp 3 \end{matrix} \quad\amp \begin{matrix}1\amp 2\amp 3\amp 4\amp 5\\ 4\amp 5\amp 1\amp 2\amp 3\\ 2\amp 3\amp 4\amp 5\amp 1\\ 5\amp 1\amp 2\amp 3\amp 4\\ 3\amp 4\amp 5\amp 1\amp 2 \end{matrix} \end{matrix} \end{equation*}

Solutions to Exercise 18.4.4

18.4.4.1 No, not every design with \(\displaystyle \lambda=1\) is is a projective plane. The condition \(\displaystyle \lambda=1\) ensures that every two points have a unique line that is incident with both of them. However, there is no requirement in a design that every two blocks have a nonempty intersection. (If every two blocks do have a nonempty intersection, then the condition \(\displaystyle \lambda=1\) does force the intersection to have exactly one point.) The condition that there exist four points such that no three are incident with a single line can also fail, but only in trivial or complete situations.

18.4.4.3 From Theorem 16.2.7, we know that there are \(\displaystyle p-1\) MOLS of order \(\displaystyle p\) whenever \(\displaystyle p\) is prime (in fact, whenever \(\displaystyle p\) is a prime power, though we did not prove this). This implies that there is a projective plane that has \(\displaystyle p+1\) points on each line whenever \(\displaystyle p\) is prime (or a prime power).

18.4.4.4 Since there are not \(\displaystyle 5\) MOLS of order \(\displaystyle 6\) (as we saw in Euler's problem), there is no projective plane that has \(\displaystyle 7\) points on each line.