SPHERICAL

TRIGONOMETRY

”T HERE IS NO ROYAL ROAD TO

GEOMETRY ”

E UCLID

BY

KELLY LYNCH
Saint Mary’s College of California
Moraga
2016
Abstract
The purpose of this paper is to derive various trigonometric formulas for spher-
ical triangles. The subject of spherical trigonometry has many navigational and astro-
nomical applications.
History
Geometry has been developing and evolving for many centuries. Its uses are
vast and continue to affect our every day lives. The study of the sphere in particular has
its own unique story, and has two major turning points. This study first began with the
push of astronomy and was developed in depth by the Greeks. There is speculation that
mathematical discoveries about the sphere were made as early as the second century,
but there is no proof for this.[1] The major transition for the understanding of spherical
geometry was the work of Menelaus of Alexandra. In his work Spherica the author
delves deeply into the properties of the sphere and the calculations related to lengths
and measures of a sphere’s arcs. For a long time the equations he discovered, such as
the measure of circumference of a sphere and the measure of arc lengths, were accepted
and no further study was needed.
The next major motivation for learning spherical trigonometry was religious
matters; the religion of Islam requires that the direction of Mecca is always known
for daily prayer. Menelaus’ findings were further developed during the Islamic En-
lightenment period. There is some debate as to the discovery of the Law of Sines for
spherical triangles. Possible sources of this discovery stem from the debate over the
two Muslim scientists, Abū Nasr or Abū ĺ-Wafā’. Who ever was responsible for the
progress in the Law of Sines allowed for a more concise proof to be developed later;
as well as leading to other theorems and identities on spherical trigonometry. Another
major name in geometry is Euclid. He made a big impact later in the third century;
though the system of spherical trigonometry does not incorporate parallel lines, Eu-
clidean geometry gave some insight to spherical behavior. In his work Elements Euclid
published equations which help lead us to the Pythagorean Theorem and the Law of
Cosines. Though mathematicians brought insight to this area of study, many influences
on spherical trigonometry also came from the field of science. Further discovery about
the behavior of arcs and angles became prominent in the late Renaissance period. John
Napier, a Scottish scientist who lived around the 17th century, was the first to work with
right spherical triangles and the basic identities of these shapes. Using Napier’s Rules,
the law of cosines for spheres was discovered.[1]

Definition 0.0.1. On a sphere, a great circle is the intersection of the sphere with a
plane passing though the center, or origin, of the sphere.[1]

Example 0.0.2. Great Circle

 The solid line illustrates a great circle which is the intersection of the plane X
and the sphere A. The plane in the 0.0.2 is the horizontal plane; however, the plane
can have any orientation that bisects the sphere. The circle created by this intersection
will have radius equal to the length of the radius of the sphere. It also follows that the
length of the circumference of the great circle will equal 2πr where r is the length of
the radius of the sphere. This is relevant because it enables us to calculate the length of
a circular segment by considering the relation between the inner angle and the radius of
the sphere. Therefore, for any sphere and any angle, the length of one arc segment will
just equal rθ where θ is the measure of the angle in radians. This is true because each
circular segment is just fraction of the entire circumference. We can show this relation
of the angle and arc length in the example below.

Example 0.0.3. General arc measure (where a is the length of the arc)

This is a general example, but we can apply specific values as in the example on
the following page.
Example 0.0.4. Measure of Arc

In this example we cut our a segment of great circle by relating an inner angle.
For this example let R = 4.69 and θ = .93759 radians, thus (4.69)(.93759) = 4.39,
which is the measure of the length of the arc. This example was created in a dynamic
mathematics software program which gave these related measurements.

The next definition required in understanding a spherical triangle is that of a

lune.

Definition 0.0.5. Lune

A lune is a part of the sphere which is captured between two great circles.[1]

This definition is relevant because it started the ability to capture shapes on a

sphere. This definition itself is not extremely significant, but it is through this shape
which we can form other shapes on the sphere. The shape of a lune can also be seen in
the next figure.

Example 0.0.6. Lune

Spherical triangles can be defined in terms of lunes.

Definition 0.0.7. Spherical Triangle

A spherical triangle is the intersection of three distinct lunes.[1]

In the figure above we can consider that there are two lunes which are the on
opposite sides of the sphere, it is natural that another lune bisecting these two will be
needed. A simpler way to think about a spherical triangle is the shape captured through
the intersection of three great circles. It is also interesting to consider that any two
unique points, which are not diametric on the spheres surface, lie on a great circle. This
makes sense because a triangle would have three vertices and therefore three different
great circles which go through two different points. Since the spherical triangle’s edges
are curved, it is clear that the equations, sides, area, and general properties will be
different from that of a planar triangle. We also know that in some way we should be
able to relate the functions that we do discover to the radius of the sphere, because it
seems natural that the formulas would change in relation to the sphere’s radius.We know
that the behavior of a spherical triangle will be very different from the planar triangles,
there are even some definitions which can illustrate this for us.

Example 0.0.8. Spherical Triangle

Definition 0.0.9. Spherical Excess is the amount by which the sum of the angles (in
the spherical plane only) exceed 180◦ .

This definition tells us about the behavior of the sphere and its edges. We know
that the length of the edges on a spherical triangle will be greater the edges on a corre-
sponding planar triangle, since they are curved. This definition allows for a spherical
triangle to have multiple right angles.
Example 0.0.10. Spherical triangle with three right angles

Definition 0.0.11. [3]

Spherical Coordinates is a coordinate system in three dimentions. The coordinate
values stated below require r to be the length of the radius to the point P on the sphere.
The value ϕ the angle between the z-axis, and the vector from the origin to point P , and
θ the angle between the x-axis, and the same vector as in the figure 0.0.12. Then we can
say the x, y, z coordinates are defined:

(r cos(θ)sin(ϕ), r sin(θ)sin(ϕ), r cos(ϕ))

Example 0.0.12. Spherical Coordinates in the xyz plane

Theorem 0.0.13. The Spherical Pythagorean Theorem[3]
For a right triangle, ABC on a sphere of radius r, with right angle at vertex C and
sides length a, b, c is defined:

c a  
b
cos = cos · cos
r r r

This equation can equivalently be written as cos(c) = cos(a)cos(b) when the

radius is 1.

Proof. Consider a spherical triangle on the surface of a sphere with radius r, and sides
a, b, c. Let vertex C in the spherical triangle be a right angle, as in the diagram below.[3]

Without loss of generality, suppose C is defined by spherical coordinates C = (0, 0, r).

Now consider the other vertices on the spherical triangle. Call one of these A and let
it be defined using spherical coordinates defined above A = (r sin rb , 0, r cos rb )

where b is equal to the measure of the arc length along the great circle between A and
b
C. Note r
is necessarily also equal the interior angle between the vector which goes
from the O to C and the vector from O to A, since the radius is equal to r. Let the
other vertex called B be defined B = (0, r sin ar , r cos ar ) where a is the arc length



 ~ = r sin b b



between B to C, by the same reasoning. Thus A r
x̂ + r cos r
ẑ and
~ = r sin a ŷ + r cos a ẑ. Now consider A
~·B
~



B r r

   
~ ~ b b a a
A · B = (r sin x̂ + r cos ẑ · r sin ŷ + r cos ẑ
r r r r
 
2 b a
= r cos cos
r r

~B ~ = r2 cos b a



Thereforesince the dot product of two vectors results in a scalar then A· r
cos r
.
~ · B|.
Now consider the magnitude of |A ~ Consider

c
~ ~
|A · B| = ||A||B||cos
r
~ =r
|A| since the radius is r
~ =r
|B| since the radius is r
c
~ · B|
|A ~ = r2 cos
r

Recall that the cosine of the angle between two vectors is just equal to the dot product
of each vector, over the magnitude of each vector. We know that the sign will not matter
since the magnitude takes the absolute value so the signs will be equal, now using this
information in relation to the equations above we can conclude that

c a  
b
cos = cos cos
r r r
Theorem 0.0.14. Law of Cosines
The Law of Cosines states that if a,b,c are the sides and A,B,C are the angles of a spher-
ical triangle, then[1]

c a    
b a b
cos = cos cos + sin sin cos(C)
r r r r r

Proof. Let there be a spherical triangle with sides denoted a, b, c. Let their opposing
angles be labeled A, B, C, where A 6= 90◦ , B 6= 90◦ , and C 6= 90◦ . Let us call the
center of the sphere to be S. Therefore we know from any single vertex, A, B, C, to the
point S is exactly equal to the radius, r. If we arbitrary let A be the largest angle(since
we can arbitrarily name any angle) then we can that there is a point somewhere between
the vertices B and C which creates a right angle, since A is the largest angle and we
must have spherical excess, it follows that creating a right angle is possible. Call this
vertex N . This partition will create two right spherical triangles. Now we know that N
some distance n from B and is the distance (a − n) from vertex C to N . There will also
be some distance we shall call y from the vertex N to point A. With this information
we can use the spherical Pythagorean equation. So now applying the equation we are
left with two equations

cos(b) = cos(n)cos(y) Equation I

cos(c) = cos(a − n)cos(y) Equation II

Now we can solve for cos(y) in equation I so therefore we have:

cos(c)
cos(y) =
cos(a − n)

Now we need to solve using the same process for equation II so we have:

cos(b)
cos(y) =
cos(n)
cos(c) cos(b)
= combining equations
cos(a − n) cos(n)
cos(c)cos(n) = cos(b)cos(a − n) cross multiply

cos(c)cos(n) = cos(b)(cos(a)cos(n) + sin(a)sin(n))

cos(c)cos(n) = cos(b)cos(a)cos(n) + cos(b)sin(a)sin(n))

sin(n)
cos(c) = cos(b)cos(a) + sin(a)sin(b) divide by cos(n)
cos(n)
cos(c) = cos(a)cos(b) + sin(a)cos(b)tan(n) Napier1 [1]

cos(c) = cos(a)cos(b) + sin(a)sin(b)cos(C)

which gives the law of cosines as desired.

Also we can put this in terms of the radius of the sphere since we know that
x
θ= R
where x is the arc length and R is the radius of the sphere so if we substitute for
the sides and angles with respect to the radius we have
1
Napier discovered the substitution for this proof that cos(b)tan(n) = sin(b)cos(C)
 c a    
b a b
cos = cos cos + sin sin cos(C)
r r r r r

By equivalence we know that

a c    
b c b
cos = cos cos + sin sin cos(A)
r r r r r

   
b a c a b
cos = cos cos + sin sin cos(B)
r r r r r

This substitution can be made in any equation since the relation of inner angles
to the arc length and r will always be consistent for that particular sphere.

Example 0.0.15. Flight Routes

Spherical trigonometry is very applicable to our every day life. One of these applications
is discovering the length and directions needed for navigation. The formulas easily
apply to the earth because the latitude and longitude lines are actually examples of great
circles. The law of cosines can be used on the example below, in which we want to find
the distance between two cities.
Given:
Earth’s radius: 6,371 km
Distance from San Francisco, California to Dublin, Ireland: 8198 km
Distance from San Francisco, California to Seattle, Washington: 1094 km
Angle between these two arcs: 0.4763342 (radians)
Now let us say we wish to find the distance from Dublin to Seattle. To find this we use
the equation below and then substitute the numerical values.

 a    
b a b
c = r arccos cos cos + sin sin cos(C)
r r r r

         
1094 8850 1094 8850
c = 6371 arccos cos cos + sin sin cos(.47633)
6371 6371 6371 6371

c = 7232 km

This value is an estimation because it assumes the earth is a perfect sphere with no dips
or irregularities which is not true. This estimation is useful, however; since it was so
quick and easy to find. Now we at least have some idea of the distance between two
places that would otherwise be impossible to directly measure.

Theorem 0.0.16. Law of Sines[4]

Where a,b,c are the sides of a spherical triangle and A,B,C are the angles opposing
these sides, the Law of Sines states the following

sin ar sin rb sin rc

= =
sin(A) sin(B) sin(C)
 ~ be the vector which
Proof. Let there be a spherical triangle ABC, with radius r. Let A
~ be the vector which goes from the
goes from the origin of the sphere to the vertex A, B
~ be the vector which goes from the origin of
origin of the sphere to the vertex B, and C
the sphere to the vertex C Now consider

C~ ×B ~
~x1 =
r2 sin ar

A~×C ~
~x2 =
r2 sin rb

B~ ×A ~
~x3 =
r2 sin rc

Each of these are unit vectors and they also must be perpendicular to the respecting
great circles, because of the right hand rule. So x1 is orthogonal to the entire plane of
sector COB. Then x2 is orthogonal to the entire plane corresponding to sector COA.
Also x3 is orthogonal to the plane lying on sector BOA.
 Consider the vector identity below.

~ × C)
(A ~ × (C
~ × B)
~ = (C
~ · (B
~ × A)
~ C)
~

First note that,

    
~ × C)
~ × (C
~ × B)
~ = 2 b a 
(A r sin ~x2 × r2 sin ~x1
r r

Thus, by we are left with

a  
~ ~ ~ ~ 4 b
(A × C) × (C × B) = −r sin sin (~x1 × ~x2 )
r r

To determine the value of (~x1 × ~x2 ), first consider that |~x1 | = |~x2 | = 1, since x1 and x2
~ = r. Next we need to consider the angle between ~x1
are unit vectors. Also note that |C|
and ~x2 , recall that each of these angles is perpendicular to its respecting great circles.
Suppose we consider these unit vectors when they are each lying exactly at point C.
Because these unit vectors are perpendicular to the whole great circle we can arbitrarily
place them at a specific point on that great circle without loss of generality.We can
arbitrarily place the unit vectors at a point because they do not depend on location, since
they are perpendicular at any place on their respecting planes. Since ~x1 is perpendicular
to the great circle corresponding to COB, then we know at vertex C, ~x1 will be exactly
tangent to the great circle created by vertices C and A. Note also that ~x2 is perpendicular
to the great circle corresponding to AOC. Therefore by similar logic ~x2 will be exactly
tangent at vertex C to the great circle which goes through vertices B and C. Therefore
these two vectors create an angle that must correspond to the angle created at vertex C.
Thus
(1)(1) sin(C) ~
(~x1 × ~x2 ) = C
r

Substituting this value in our previous equation we get,

a  
~ ~ ~ ~ 3 b ~
(A × C) × (C × B) = −r sin sin sin(C)C
r r

So therefore by the original vector identity we have

a  
~ ~ ~ 3 b
C · (B × A) = −r sin sin sin(C)
r r

From this the following is equivalent,

 
~ · (C
~ × B)
~ = −r3 sin b c
A sin sin(A)
r r

~ = −r3 sin c sin a sin(B)

   
~ · (A
B ~ × C)
r r
Also note that the triple product of vectors corresponds to the determinant. From
properites in linear algebra we know that determinants do not change when two pairs of
rows are swapped, which means it is invariant. Therefore, since we are only swapping
two pairs of vectors; we know that the equations above all equivalent.

a    
b b c c a
sin sin sin(C) = sin sin sin(A) = sin sin sin(B)
r r r r r r

a b c




Then dividing this whole expression by sin r
sin r
sin r
, we have

sin(C) sin(A) sin(B)

c

= a

=
sin rb


sin r sin r

which equivalently results in

sin ar sin rb sin rc

= =
sin(A) sin(B) sin(C)

as desired.

Now that we have the Pythagorean Theorem, Law of Cosines, and the Law of
Sines for a sphere, we can compare these to their planar counterparts. We can consider
the following comparison.

Spherical (Let r=1) Cosines:

Pythagorean Theorem:
cos(c) = cos(a)cos(b)+sin(a)sin(b)cos(C)

cos(c) = cos(a)cos(b)
Sines: Planar
Pythagorean Theorem:
sin(a) sin(b) sin(c)
= =
sin(A) sin(B) sin(C)
c 2 = a2 + b 2

Cosines:

c2 = a2 + b2 − 2abcos(C)

Sines:

a b c
= =
sin(A) sin(B) sin(C)

Now it is interesting to see that the Law of Sines is quite similar for both pla-
nar and spherical. In contrast the Pythagorean Theorem and Law of Cosines are quite
different.It is also interesting to note that in both planar and spherical the Pythagorean
Theorem is a limit of the Law of Cosines. For the spherical equations this seems natural
since we must use the Pythagorean Theorem in order to prove the Law of Cosines; as
shown in the proof above. It is important that we recognize the fact that spherical trian-
gles and planar triangles are very different and the comparison of these equations help
to show us the differences.

Example 0.0.17. Law of Sines Application

There are many examples we can use for the application for the Law of Sines.
To keep variety I will including an example which relates to astronomy. Suppose that
we have three lunar rovers, call them A,B, and C respectively. Now suppose we wish
to find the distance between two of these such rovers. We can use the following given
information and the Law of Sines to discover a particular length.
Radius of moon:1737 km
Distance between rover C and B (length a): 10 km
3π
Angle created at rover B: 4
π
Angle created at rover A: 6

How far are rovers A and C (length of b)?

sin( ar ) sin( rb )
=
sin(A) sin(B)

Now we can substitute the numerical values and solve for b so we have,

sin( 3π 10
 
4
) sin( 1737 )
b = 1737 arcsin π
sin( 6 )

b = 14.22 km

Thus the distance between rover A and rover C is around 14 km. This value does
not take into account any variation, such as dips or holes on the moon’s surface so it is
best used as a rough estimation. Again it was quite simple to find and gives some idea
of how far these lunar rovers are from eachother.

Theses are just a few of the many examples spherical trigonometry has in our
world. It is very useful in the study of astronomy and in any navigational problems on
our Earth. Having the ability to relate the spherical Phythagoream Theorem, Law of
Cosines, and Law of Sines to the radius of a sphere allows us use these equations on
spheres of different sizes. This also gives us more freedom in the use of these equations
as a whole, over various disciplines such as math, physics, and astromony. Deriving the
spherical trigonometric formulas in terms of the radius, was influential in history and its
applications and uses continue to the modern day.
Bibliography

[1] Brummelen, Glen Van. Heavenly Mathematics: The Forgotten Art of Spherical
Trigonometry. Princeton: Princeton UP, 2013. Print.

[2] Clark, C. E. Proofs of the Fundamental Theorems of Spherical Trigonometry. Math-

ematics Magazine 21.3 (1948): 151. Web.

[3] McCleary, John. Geometry from a Differentiable Viewpoint. Cambridge: Cam-

bridge UP, 1994. Print.

[4] Wilson, P. M. H. Curved Spaces: From Classical Geometries to Elementary Differ-

ential Geometry. Cambridge: Cambridge UP, 2008. Print.

22