Difference Between Euclidean and Spherical Trigonometry

1

Non-Euclidean geometry is geometry that is not based on the postulates of Euclidean geometry. The five postulates of Euclidean geometry are:
1. Two points determine one line segment.
2. A line segment can be extended infinitely.
3. A center and radius determine a circle.
4. All right angles are congruent.
5. Given a line and a point not on the line, there exists exactly one line containing the given point parallel to the given line.
The fifth postulate is sometimes called the parallel postulate. It determines the curvature of the geometry’s space. If there is one line parallel to the given line (like in Euclidean geometry), it has no curvature.
If there are at least two lines parallel to the given line, it has a negative curvature. If there are no lines parallel to the given line, it has a positive curvature. The most important non-Euclidean geometries are hyperbolic geometry and spherical geometry.
Hyperbolic geometry is the geometry on a hyperbolic surface. A hyperbolic surface has a negative curvature. Thus, the fifth postulate of hyperbolic geometry is that there are at least two lines parallel to the given line through the given point.

2

Spherical geometry is the geometry on the surface of a sphere. The five postulates of spherical geometry are:
1. Two points determine one line segment, unless the points are antipodal (the endpoints of a diameter of the sphere), in which case they determine an infinite number of line segments.
2. A line segment can be extended until its length equals the circumference of the sphere.
3. A center and a radius with length less than or equal to πr where r is the radius of the sphere determine a circle.
4. All right angles are congruent.
5. Given a line and a point not on the line, there are no lines through the given point parallel to the given line.
A line segment on a sphere is the shortest distance between two points on the sphere. This distance is the smaller section of arc formed by the two points on a great circle. A great circle is a circle whose center is the center of the sphere and whose radius is the radius of the sphere.
In spherical geometry, a triangle is the section of a sphere bounded by the arcs of three great circles. There are two sections formed in this way. One is the section whose interior angles each have a measure less than 180°. The other is the section whose interior angles each have a

3

measure greater than 180°. I will be working only with the first triangle, and will regard it as the spherical triangle formed by the arcs of three great circles. Also, each side must have a length less than or equal to half the circumference of the sphere because otherwise the side would not fit the definition of a line segment.
I will focus on the following problem: For a triangle ABC — where a, b, and c are the lengths of the sides opposite angles A, B, and C, respectively —, given a, b, and the measure of angle C, what is the relationship between c where ∆ABC is in a Euclidean plane and c where
∆ABC is on the surface of a sphere, and what is the relationship between the area of ∆ABC in a Euclidean plane and ∆ABC on the surface of a sphere? In other words, find

c Euclidean
Area Euclidean and . The values of these c spherical
Area spherical

ratios will vary depending on the size and shape of the triangles, as well as how large they are relative to the sphere. I will explore these ratios for a variety of circumstances.
To do this, I will use spherical trigonometry. Because ∆ABC is not necessarily right, I will use trigonometry for oblique triangles. The key formulae for solving this problem are the law of cosines for sides and the law of sines. The derivations of them are much like the derivations for the

4

law of cosines and law of sines in Euclidean trigonometry. The first step is to derive a trigonometry for right triangles.
The measure of the angle formed by the intersection of two great circles is defined to be equal to the measure of the angle formed by the two lines tangent to each great circle at the point of intersection. It is also equal to the measure of the dihedral angle formed by the planes of the great circles. (The measure of a dihedral angle is the measure of the angle formed by the two lines of intersection made by the two planes forming the dihedral angle with a plane perpendicular to the two planes’ line of intersection.) This is shown in the following proof.
Call the point of intersection of the two great circles A and the center of the sphere O. Let lines l and m each be tangent at A to one of the great circles that intersect at A. l ⊥OA and m ⊥ OA because l and m are each tangent to a circle that OA is a radius of (and a line tangent to a circle is perpendicular to the radius of the circle that intersects it).
Because of this, the measure of the dihedral angle formed by planes l-O and m-O is equal to the measure of the angle formed by l and m, and therefore equal to the measure of the spherical angle formed by the intersection of the great circles.

5

1

In the picture above, O is the center of the sphere. ∆ABC is a right spherical triangle with the right angle at C.
Construct plane DEF through any point E on OB such that plane DEF ⊥
OA. Let D be on OA, and F be on OC. DE ⊥ OA and DF ⊥ OA because DEF ⊥
OA and DE and DF are on DEF. Then ∆ODF is a right triangle with its right angle at D because DF ⊥ OA, and ∆ODE is a right triangle with its right angle at D because DE ⊥ OA. DEF ⊥ OAC because DEF ⊥ OA and OA is on
OAC. If two planes intersect and are each perpendicular to a third plane then their line of intersection is perpendicular to the third plane; so EF ⊥
OAC because
BCO ⊥ OCA and DEF ⊥ OCA. OC is on OCA, therefore EF ⊥ OC, so ∆OFE is right. Also, DF is on OCA, therefore EF ⊥ DF, so ∆DFE is right.
For each part of spherical triangle ABC, there is an angle whose measure equals the measure of the part (I will be expressing all angle and

6

arc measurements in radians). Each part and corresponding angle are: m∠A = m∠EDF

a = m∠FOE

m∠B = m∠DEF

b = m∠DOF

m∠C = m∠DFE

c = m∠DOE

Because of this, sin a = sin m∠FOE =

FE
FE ED
=
= sin A sin c.
OE
ED OE

Using similar logic and constructing a plane perpendicular to OB instead of OA, six more formulas are obtained for a total of seven. Using these seven formulas an additional three formulas can be derived to make a total of ten formulas. These ten formulas are: sin a = sin A sin c

sin b = sin c sin B

tan a = sin b tan A

tan b = sin a tan B

tan a = tan c cos B

tan b = tan c cos A

cos c = cos a cos b

cos c = cot A cot B

cos A = cos a cos B

cos B = cos b sin A

2

To derive the law of sines and the law of cosines for sides, the construction of an altitude must be made. Construct the altitude from vertex C to side c and extend if necessary. The two cases (not extending side c and extending side c) a are depicted here:

7

3

Case I: c does not need to be extended.

Case II: c does need to be extended.

Law of Sines
The identities of right spherical triangles listed before yield: sin h = sin a sin B, and sin h = sin b sin A
Substituting for sin h, sin a sin B = sin b sin A. Dividing both sides by sin A sin B,

sin a sin b
=
. sin A sin B
If the altitude h were constructed from vertex B to side b, the derived identity would be:

sin a sin c
=
sin A sin C
By transitivity,

sin a sin b sin c
=
=
.
sin A sin B sin C

4

Law of Cosines for Sides
In both cases, ∆ADC is right with ∠D being the right angle. Applying

8

identities for right spherical triangles, cos a = cos h cos (c-m) for the first case, and cos a = cos h cos (m-c) for the second case.
Because cos (c-m) = cos (m-c), cos h cos (c-m) = cos h cos (m-c).
Therefore, the first case can be used to derive the formula for both cases.
Because cos (α-β) = cos α cos β + sin α sin β, cos h cos (c-m) = cos h (cos c cos m + sin c sin m)
Applying the identities for right spherical triangles to ∆ADC, cos b = cos h cos m (dividing both sides by cos h gives cos m =

cos b
),
cos h

sin m = tan h cot A, and sin h = sin b sin A.
Substituting for cos m and sin m, ä å å cos h (cos c cos m + sin c sin m ) = cos h åcos c å ã

ë cos b ì ì
+ sin c tan h cot A ì ì cos h í Distributing, ä ë cos b å ì å ì cos h åcos c
+ sin c tan h cot A ì = cos c cos b + sin c sin h cot A å ì cos h ã í
Substituting sin h = sin b sin A and using the fact that sin A cot A = cos A, cos c cos b + sin c sin h cot A = cos c cos b + sin c sin b cos A.

9

Therefore, by transitivity, cos a = cos c cos b + sin c sin b cos A.
By constructing the altitude h to sides a and b instead of c, these three forms of the law of cosines for sides can be derived: cos a = cos c cos b + sin c sin b cos A cos b = cos a cos c + sin a sin c cos B cos c = cos b cos a + sin b sin a cos C

5

For an example, I will use the triangle formed on the Earth by New
York, Moscow and the North Pole. The distance between New York and
Moscow will be unknown.
The Earth can be treated as a sphere with a radius of 3959mi. Such a sphere is called the terrestrial sphere. On the terrestrial sphere, the equator is the great circle that determines a plane perpendicular to the line drawn between the North Pole and the South Pole. A meridian is a great circle that intersects both poles. The prime meridian is the meridian that Greenwich, England intersects. In the following picture, the North Pole is at point N and the center of the sphere is at point O.

10

The longitude of a point P on the sphere is the measure of the arc intercepted on the equator by the meridian that P is on and the prime meridian. In the above picture, it is the measure of arc EF (which equals m∠EOF). Longitude is positive when P is West of the prime meridian and negative for when P is East of the prime meridian. The latitude of a point P on the sphere is the measure of the arc intercepted on the meridian that P is on by P and the equator. In the above picture, it is the measure of arc PE
(which equals m∠POE). Latitude is positive when P is North of the equator and negative when P is South of the equator. The colattitude of a point P is π/2 - the latitude of point P. This is the measure of the arc intercepted on the meridian P is on by P and the North Pole (the measure of arc PN).

11

In the picture above, N is New York, M is Moscow, and P is the North
Pole. The distance between the North Pole and New York is New York’s colattitude (NP), and the distance between the North Pole and Moscow is
Moscow’s colattitude (MP). The difference in longitude between New York and Moscow is the arc intercepted on the equator by the meridians they are on. In the above picture, this is arc EF. The measure of arc EF = m∠EOF.
Because the plane the equator is in is perpendicular to the line determined by the North Pole and South Pole, m∠EOF equals the measure of the dihedral angle formed by plane NEO and plane MFO (namely, N-POM). Because m∠EOF equals the measure of dihedral angle N-PO-M and the measure of N-PO-M is the measure of spherical angle NPM, m∠EOF equals the measure of spherical angle NPM. m∠EOF equals the measure of arc EF, which is the difference in longitude between New York and Moscow.

12

Therefore, the measure of the angle formed by New York and Moscow with North Pole (the North Pole as the vertex) is the difference in longitude between the two.
New York’s longitude is 37π/90, and its latitude is 2443π/10800.
Moscow’s longitude is -1127π/5400 and latitude is 223π/720. n = colattitude of N = π/2 - 2443π/10800 = 2957π/10800. m = colattitude of M = π/2 - 223π/720 = 137π/720.
P = difference in longitude = 37π/90 + 1127π/5400 = 3347π/5400.
By the law of cosines for sides, cos NM = cos m cos n + sin m sin n cos P = .382398175, so
NM = Arccos .382398175 = 1.178405986 radians. To get the result in miles, this is multiplied by the radius (3959mi) to get 4665.309297mi. If one used Euclidean trigonometry to calculate this,

NM 2 = m 2 + n 2 − 2mncos P = 23122014.49, so
NM =

23122014.49 = 4808. 535587 mi.
The result from using Euclidean trigonometry is significantly larger

than the result from spherical trigonometry. This is because the given lengths wrap along the sphere, making the distance between them smaller than if they were straight.

13

The problem I presented at the beginning of the paper is: For a triangle ABC — where a, b, and c are the lengths of the sides opposite angles A, B, and C, respectively —, given a, b, and the measure of angle C, what is the relationship between c where ∆ABC is in a Euclidean plane, and c where ∆ABC is on the surface of a sphere and what is the relationship between the area of ∆ABC in a Euclidean plane and ∆ABC on the surface of a sphere? In other words, find

c Euclidean
Area Euclidean and . The triangles in this c spherical
Area spherical

problem are illustrated here:

Euclidian

The first part of the problem is to find

Spherical

c Euclidean
. The first part of c spherical

finding this is to find the Euclidean c and spherical c independently.

14

Euclidean

c 2 = a 2 + b 2 − 2abcos C àc =

a2 + b 2 − 2abcos C
Spherical

I will use the constant r as the radius of the sphere. To apply law of cosines for sides, everything must be expressed in arc length.

length arc
=
2πr
2π
arc =

length r Applying the law of cosines for sides,

c a b
= cos cos r r r c ä a å
= Arccos å cos cos å r r ã ä a å c = r Arccos å cos cos å r ã cos a b sin r r a + sin sin r a + sin sin r

+ sin b r b r

cos C ë b ì cos C ì ì r í ë b ì cos C ì ì r í Therefore, c Euclidean
=
c spherical

a2 + b 2 − 2abcos C ë ä a b a b å ì r Arccos å cos cos + sin sin cos C ì å ì r r r r ã í

The second part of the problem is to find

Area Euclidean
. The first part of
Area spherical

15

finding this is to find each area independently.
Euclidean
Put ∆ABC in a coordinate plane, with C at the origin.

sin C =

y
1
. à bsin C = y . Area = ( 2 )(base)(height), so Area = b 1
2

ay .

Substituting for y,
Area =

1
2

absin C

Spherical

Using to law of sines to get A and B,

sin A sin B sin C
=
= a b c . Therefore, sin sin sin r r r

ë ë ä a ä b å ì å ì åsin åsin sin C ì sin C ì å ì å ì å ì ì å ì å ì r r å ì and B = Arcsin å å ì å ì å ì
A = Arcsin å ì å ì å ì å ì å ì å ì å c ì å c ì å ì å ì å ì å ì sin sin å ì å ì r r ã í ã í The area of a spherical triangle is equal to r 2 E , where E is the angle excess of the triangle.6 The angle excess of a triangle is the amount the

16

sum of the measures of its angles is over π (the sum of the angles - π). For example, the 90°-90°-90° triangle (whose angles are all 90°, or π/2 radians) has an area 1/8 of the surface area of the sphere. Using the angle

r2 excess formula, its area =
. The surface area of a sphere is 4π r 2 , so the
2

result from the angle excess formula is true.
In the original problem, the angle excess formula gives:
Area = r 2 á + B + C −
A
ä å å å å å 2å
Area = r å Arcsin å å å å å å å ã

. é Substituting for A and B,

ë ë ä a ä b å ì å ì åsin åsin sin C ì sin C ì å ì å ì å ì ì å ì å ì r r å ì + Arcsin å å ì + C − å ì å ì å ì å ì å ì å ì å ì å ì å c ì å c ì å ì å ì å ì å ì sin sin å ì å ì r r ã í ã í

ë ì ì ì ì ì ì ì ì ì ì ì ì ì ì í Therefore,
Area Euclidean
=
Area spherical

1
2

ä å å å å å 2å r å Arcsin å å å å å å å ã

absin C

ë ë ä a ä b å ì å ì åsin åsin sin C ì sin C ì å ì å ì å ì ì å ì å ì r r å ì + Arcsin å å ì + C − å ì å ì å ì å ì å ì å ì å ì å ì å c ì å c ì å ì å ì å ì å ì sin sin å ì å ì r r ã í ã í

ë ì ì ì ì ì ì ì ì ì ì ì ì ì ì í I will explore these ratios for the special case where a=b=k on the unit sphere (where the radius, r, equals one). Because the only triangles I am dealing with are those whose angles each have a measure less than π

17

and greater than 0, 0 < C < π. Also, 0 < k < π because otherwise k would not fit the definition of a line segment. Below are graphs of the

ratio as a function of C (

c Euclidean c spherical

c Euclidean is on the y-axis and C is on the x-axis). c spherical

There are two graphs: one where k=π/2 and another where k=1.2.

k = π/2 y=1 k=1.
2

As k increases,

y=1

c Euclidean increases (no matter what the value of C is). c spherical

18

This implies that the spherical c becomes increasingly smaller than the
Euclidean c as more of the sphere is covered by sides a and b.
The limit of

c Euclidean as k approaches zero is 1. This implies that for c spherical

small values of k, the difference between Euclidean c and spherical c is negligible. An example of this is when one deals with local triangles on the
Earth. If one is dealing with triangles that are small relative to the size of the Earth (these are about all of the triangles encountered in every-day life), one can use Euclidean trigonometry and get results just about as accurate as one would have gotten had they used spherical trigonometry.
The limit of

c Euclidean as C approaches π is 1. This is because the length c spherical

of side c is getting closer to a+b (which always has the same value in the
Euclidean triangle as in the spherical triangle).
When k > π/2, there is a “liftoff” of the entire graph. The graph of c Euclidean where k=2.4 (a number greater than π/2) is presented below: c spherical

19

k = 2.4

y=1

This liftoff is due to a wrap-around effect. Wrap-around occurs when k is greater than one quarter of the circumference of the sphere. When this happens, the length of side c decreases as k increases because sides a and and b are wrapping around the place where c is maximized. Thus, the maximum value for c is when k=π/2. My proof of this is presented here:

c = Arccos ácos 2 k + sin 2 k cos C é
Differentiating with respect to k yields: ä dc
−1
å å = å å å å 2 dk cos2 k + sin 2 k cosC é ã 1− á
Factoring out 2cos k sin k ,

ë ì ì á 2cos k sin k + 2sin k cos k cos C é ì − ì ì ì í

20

dc
=
dk

−1
1 − á k + sin k cos C é cos 2

2

dc
=
dk

2

á cos k sin k á C − 1 é
2
cos é − 2cos k sin k á C − 1 é cos 1 − á 2 k + sin 2 k cos C é cos 2

The critical points are at the values of k that make the derivative either equal to zero or undefined. The derivative equals zero when its numerator equals zero, and it is undefined when the denominator equals zero. Thus, the values of k that are critical points are values that make the numerator equal to zero or make the denominator equal to zero.
Case I: the numerator equals zero
Setting the numerator equal to zero,

− 2cos k sin k á C − 1 é = 0 cos Both cos C - 1 and -2 are constants, so both sides of the equation can be divided by − 2 á C − 1 éwithout losing a root. Executing this division, cos cos k sin k = 0
If a product of two numbers equals zero, than one or both must equal zero. Therefore,

cos k = 0 or sin k = 0 π àk = or k = 0 or k = π
2

21

Case II: the denominator equals zero
Setting the denominator equal to zero,

1 − á 2 k + sin 2 k cos C é = 0 cos 2

Squaring both sides, adding á 2 k + sin 2 k cos C é to both sides, and then cos 2

taking the square root of both sides,

cos2 k + sin 2 k cos C = " 1
Substituting 1 − cos 2 k for sin 2 k ,

cos2 k + á − cos 2 k é C = " 1
1
cos
Distributing cos C,

cos2 k + cos C − cos 2 k cos C = " 1
Factoring out cos2 k ,

á − cos C é 2 k + cos C = " 1
1
cos
Subtracting cos C from both sides and dividing both sides by (1 - cos C),

cos2 k =

" 1 − cos C
1 − cos C ä å å à k = Arccos å" å ã

" 1 − cos C
1 − cos C

ë ì ì ì ì í Because the radicand must be positive and |cos C| < 1, the -1 of the " 1 must be discarded. Therefore,

22

ä

å å k = Arccos å" å ã

1 − cos C
1 − cos C

ë ì ì ì ì í à k = Arccos á 1 é = Arccos á 1 é
"
" à k = 0 or k = π
The set of all the critical points is the union of the points from Case I
(numerator = 0) and the points from Case II (denominator = 0).
Therefore, the critical points are:

k =

π
, k = 0 , and k = π
2

To find which are relative maxima, I will make a sign diagram of

dc
.
dk

dc is reprinted here: dk dc
=
dk

− 2cos k sin k á C − 1 é cos 1 − á 2 k + sin 2 k cos C é cos 2

(cos C - 1) is always negative because |cos C| < 1. The denominator is always greater than or equal to zero because it is a positive square root.2 is always negative, so the sign of

dc when k is not a critical point is the dk sign of cos k sin k .
When 0 < k <

π dc , cos k and sin k are both positive, so is positive.
2
dk

23

When

π dc < k < π , cos k is negative and sin k is positive, so is 2 dk negative when

π
< k < π . The sign diagram is drawn below:
2

++++++++++ - - - - - - - - - - -

0

π

dc π dk

2
Thus, c is increasing over the interval (0, π/2) and decreasing over the interval (π/2, π). Therefore, there is a relative maximum at k =

Within the domain there is no other relative maximum, so k =

π
.
2

π is the
2

absolute maximum for c.
Below are two graphs of

Area Euclidean
Area Euclidean as a function of C. is on
Area spherical
Area spherical

the y-axis, and C is on the x-axis. In one graph, k=1. In the other, k=π/2.

24

k=1 y=1 k=π/2 y=1 The limit as k approaches zero of

Area Euclidean equals one. As k increases, the
Area spherical

y-intercept increases. Also as k increases, the limit as C approaches π of
Area Euclidean decreases, and reaches a minimum of zero at k = π/2. When
Area spherical

k > π/2, the wrap-around effect creates a liftoff of the graph. A graph where k > π/2 is presented here:

25

k=2

y=1

In the amount of time I was given to complete this paper, I was unable to accomplish several things. I could not explore the

Area Euclidean
Area spherical

ratio for isosceles triangles as much as I wanted. Also, I did not have enough time to explore the two ratios for non-isosceles triangles. I believe wraparound affects the ratios for non-isosceles triangles when the sum a+b is greater than πr/2 (π/2 on the unit sphere). In addition, I was unable to explore the ratios when the radius of the sphere is a constant r.
However, the work I did where the radius equals one can apply to all

26

spheres; the units of length can always be defined such that the length of the radius of the sphere equals one. For example, on Earth one can measure everything in terms of Earth radii.

27

NOTES

1. Picture from Paul R. Rider, Plane and Spherical Trigonometry, New York: The
Macmillan Company, p. 201

2. Ten formulas from Rider, Plane and Spherical Trigonometry p. 203

3. Both pictures from Rider, Plane and Spherical Trigonometry p. 212

4. Derivation adapted from Rider, Plane and Spherical Trigonometry pp. 211-212.

5. Derivation adapted from Rider, Plane and Spherical Trigonometry pp. 212-213

6. Angle excess formula adapted from Rider, Plane and Spherical Trigonometry p. 200