MATH 324 Summer 2006

Elementary Number Theory

Notes on Pick’s Theorem

Department of Mathematical and Statistical Sciences

University of Alberta

Pick’s Theorem

The forest industry is concerned with the calculation of the area inside a polygonal region drawn to scale
from field data obtained from a stand of timber. The standard method is to overlay a scale drawing with a
transparency on which a square dot pattern is drawn. The area inside the polygon is computed by counting
all of the dots fully inside the polygon, and adding half of the number of dots which fall on the bounding
edges of the polygon. The method used is called Pick’s formula.

The points in the plane with integer coordinates are called lattice points. A simple closed polygon in the
plane is a lattice polygon if all of its vertices are lattice points. Every lattice point on a lattice polygon is
considered a vertex, even when the angle at the vertex is straight.

The simplest lattice polygon is a fundamental triangle, a triangle which contains no interior lattice points
and whose boundary contains only the three lattice points at its three vertices.

A triangulation of a lattice polygon is a partitioning of the interior of the polygon into fundamental triangles.
The figure below shows the triangulation of a lattice polygon into fundamental triangles.

We will show the following:

(a) The measures of the interior angles at the vertices of a simple polygon with B vertices sum to
(B − 2) · 180◦ .

(b) If (0, 0), (a, b), (c, d) are the coordinates of the vertices of a triangle T in the plane, then the area of T
is
1
area(T ) = |ad − bc|.
2
 (c) Every triangulation of a lattice polygon with I interior lattice points and B boundary lattice points
requires

F = 2I +B −2
fundamental triangles.
(d) If T is a lattice triangle which is a fundamental triangle, then
1
area(T ) = .
2

Using the above we will prove Pick’s Theorem, if a lattice polygon has B lattice points on its boundary and
encloses I lattice points in its interior, then its area A is given by
1
A = I + B − 1.
2

Solution:

(a) We will give an easy induction argument to show that a convex polygon with B sides can be divided
into triangles by drawing diagonals that do not intersect and that no matter how this is done, the
number of diagonals is always B − 3 and the number of triangles is always B − 2. The proof is by
induction on the number of vertices B.

Base Case : When B = 4, then the number of diagonals needed is B − 3 = 1 and the number of
triangles formed is B − 2 = 2.

Inductive Step : Now suppose that the result is true for all convex poygons with fewer than B vertices,
and consider a polygon with B vertices and divide it into two nonoverlapping convex polygons with n
vertices and m vertices, where B = n + m − 2.

n vertices

m vertices

By the induction hypothesis, we can triangulate the convex n–gon with n − 3 diagonals to form n − 2
triangles, and we can triangulate the convex m–gon with m − 3 diagonals to form m − 2 triangles.
Therefore, the convex B–gon can be triangulated with
n−3+m−3+1=n+m−2−3=B−3
diagonals (this includes the edge shared by the n–gon and the m–gon), into
n−2+m−2=n+m−2−2=B−2
triangles. Therefore, by the principle of mathematical induction, the result is true for all convex B-
gons. If the polygon is simple, but no longer convex, an argument similar to that above shows that
the diagonals of the polygon with B vertices will divide it into B − 2 triangles, each containing 180 ◦ .
Therefore, the measure of the interior angles at the vertices of a simple polygon with B vertices sum
to (B − 2) · 180◦.
(b) If the coordinates of the vertices of a triangle T in the plane are (0, 0), (a, b), (c, d), we may assume
without loss of generality that b < d. We consider separately the three cases a = c, 0 < c < a and
0 < a < c as shown in the figure.

y y y

(c,d) (c,d)

(c,d)
T (a,b)
T
T (a,b) (a,b)

(0,0) x (0,0) x (0,0) x

case i. If a = c, then clearly,

1 1 1 1
area(T ) = ad − ab = (ad − ab) = |ad − bc|.
2 2 2 2

case ii. If 0 < c < a, then from the figure we have

1 1 1 1 1
area(T ) = cd + (a − c)b + (a − c)(d − b) − ab = (ad − bc) = |ad − bc|.
2 2 2 2 2

case iii. If 0 < a < c, then from the figure we have

1 1 1 1 1
area(T ) = cd − ab − (c − a)(d − b) − b(c − a) = (ad − bc) = |ad − bc|.
2 2 2 2 2

(c) We will show that all triangulations of a simple lattice polygon use the same number of triangles.

Suppose that P is a simple lattice polygon with I interior lattice points and B boundary lattice points,
and P has been triangulated into F fundamental triangles. The measures of the three angles of each
triangle sum to 180◦ , so that the measures of the angles in all F of the triangles sum to F · 180◦ . We
can also get this sum by adding 360◦ for each interior lattice point I, and the sum of the angles at the
B boundary points, namely, (B − 2) · 180◦ . Therefore,

F · 180◦ = I · 360◦ + (B − 2) · 180◦ ,

which gives F = 2I + B − 2, and this is the same for any triangulation of P by fundamental triangles.

(d) Note that if T is a lattice triangle with vertices (0, 0), (a, b), (c, d), then since a, b, c, d are all integers,
we must have
1
area(T ) ≥ .
2
On the other hand, suppose T1 is a fundamental triangle, and that T1 is included inside a sufficiently
large lattice square S, whose sides are parallel to the coordinate axes. If S measures n × n, then S
contains n2 unit squares. Since each unit square can be subdivided into two fundamental triangles,
it is clear that S can be triangulated into 2n2 fundamental triangles. From (c) it follows that every
triangulation of S into fundamental triangles requires 2n2 fundamental triangles.
 In particular, consider a triangulation of S that contains T1 , so that S is partitioned into the funda-
mental triangles T1 , T2 , . . . , T2n2 , then

area(S) = n2 = area(T1 ) + area(T2 ) + · · · + area(T2n2 )

1 1 1
≥ + +···+
2
| 2 {z 2}
2n2 terms
1

= 2n2
2
= n2 .

1
Therefore, if any Ti has area(Ti ) > , then we would have strict inquality in the above. Thus,
2
1
area(Ti ) = for all i = 1, 2, . . . , 2n2 . Since T1 was an arbitrary fundamental triangle, then we are
2
done.

Theorem (Pick): If a lattice polygon has B lattice points on its boundary and encloses I lattice points in
its interior, then its area A is given by
1
A = I + B − 1.
2

Proof. Every simple lattice polygon P with I interior lattice points and B boundary lattice points can be
1
triangulated with F = 2I + B − 2 fundamental triangles, each with area , therefore,
2
1 1
area(P ) = A = F = I + B − 1.
2 2