Mathematics for Celestial Navigation

Richard LAO
Port Angeles, Washington, U.S.A.
Version: 2018 January 25

Abstract
The equations of spherical trigonometry are derived via three dimensional
rotation matrices. These include the spherical law of sines, the spherical law of
cosines and the second spherical law of cosines. Versions of these with appropri-
ate symbols and aliases are also provided for those typically used in the practice
of celestial navigation. In these derivations, surface angles, e.g., azimuth and
longitude di¤erence, are unrestricted, and not limited to 180 degrees.
Additional rotation matrices and derivations are considered which yield
further equations of spherical trigonometry. Also addressed are derivations of
"Ogura’s Method " and "Ageton’s Method", which methods are used to create
short-method tables for celestial navigation.
It is this author’s opinion that in any book or paper concerned with three-
dimensional geometry, visualization is paramount; consequently, an abundance
of …gures, carefully drawn, is provided for the reader to better visualize the
positions, orientations and angles of the various lines related to the three-
dimensional object. 44 pages, 4MB. RicLAO.
Orcid Identi…er: https://orcid.org/0000-0003-2575-7803.

1 Celestial Navigation
Consider a model of the earth with a Cartesian coordinate system and an embedded
spherical coordinate system. The origin of coordinates is at the center of the earth
and the x-axis points through the meridian of Greenwich (England). This spherical
coordinate system is referred to as the celestial equator system of coordinates, also
know as the equinoctial system. Initially, all angles are measured in standard math-
ematical format; for example, the (longitude) angles have positive values measured
toward the east from the x-axis.
Initially in this paper we will measure all angles in this standard mathematical
format, that is, in the sense that a right-hand screw would turn were it to advance
from the south pole to the north pole. In the Nautical Almanac most angles are

1
tabulated westerly, in keeping with the direction of apparent travel of the sun over a
point on the surface of the earth.
We will be using two sets of symbols for angles (aliases of one to the other), one
of which is frequently used in the practice of celestial navigation. In this paper, for
economy of space in …gures, the symbol P will often be employed instead of GP
to represent the geographical position of an observed celestial body. Likewise, the
symbol M will be used to represent the position of the observer, e.g., the assumed
position AP or dead reckon position DR. The reader should note that in works of
other authors the symbol M oftentimes represents the GP instead of the position of
the observer.

Spherical Coordinate Angle Symbols and Aliases

= B = co-altitude of the celestial body P.
= A = co-declination of the celestial body P.
= A = east longitude of the celestial body P.
A = B = co-azimuth of the celestial body P relative to observer M.
= (no alias) co-latitude of the observer M.
= = east longitude of the observer M.
l= (or )= , di¤erence in east longitudes.

The celestial equator system of coordinates.

Angles A = , = and are measured in customary mathematical format;
that is, they are positive when measured in an easterly direction and expressed as
east longitudes. For westerly longitudes, is negative.
The symbol will also be used to generically represent an easterly longitude of
any other point speci…ed on the celestial sphere in the text and understood in the
context of that text.
Longitude can also be measured in customary navigational format; that is, positive
when measured in a westerly direction.
I call the rotational motion of the earth boreal motion, which usage I adopted
from Skilling1 . As Skilling notes, the word boreal is derived from the rotation of the
earth and signi…es a northerly direction compared to the rotation of the earth. An
analogous "right-hand screw", corresponding to this motion, would be driven along
the polar axis from the South pole up through the North pole.

1
Skilling, Hugh Hildreth, Fundamentals of Electric Waves, 2nd edition, 1948, page 87, reprinted
by Robert E. Krieger Publishing Company, Inc., 1974. ISBN 0-88275-180-8

2
 Figure 1: the Celestial Equator System of Coordinates.

where = Right Ascension (RA) of P

A = = Co-declination of P
A = = +
and the symbol (the "ram’s horns") is known as the …rst point of Aries.
Remember that the ”right hand screw rotates to the EAST”. is the longitude
of this point, the …rst point of Aries or the vernal (spring) equinox, measured easterly
in the equatorial plane from the Greenwich meridian.

Now consider the horizon system of coordinates.

R? is the distance from the center of the earth to the observed celestial body P .
Suppose that Robs is the distance from the observer M on the surface of the earth
to the same celestial body. Since these distances are extremely large compared to
the radius of the earth, for computation purposes we may regard these two distances
as equal to one another. Furthermore, we may regard spherical coordinate angles
of the celestial body as equal to one another whether measured from the center of
the earth or from the observer’s position on the surface. Let R be the radius of the
earth, considered constant in this paper. The geographical position GP of a celestial
body is the the point on the earth’s surface directly below the celestial body, that is,
the point of intersection with the earth’s surface of a line through the body and the
center of the earth.

3
 Figure 2: The Horizon System of Coordinates.

In the practice of celestial navigation, angles are usually measured in degrees,

minutes and seconds of arc rather than in radians. We shall adhere to that convention.
Furthermore, since the apparent motion of the sun in the sky relative to an observer
on the surface of the earth is westward, navigational angles are usually measured
westward.

The Greenwich Hour Angle GHA(P ) = 360 (1)

For example, GHA is the longitude of the sun measured westerly. The symbol
W will also be used to generically represent a westerly longitude of any other point
speci…ed on the celestial sphere in the text and understood in the context of that
text.
Whether we are expressing all longitudes in the easterly direction (standard
mathematical angle format) or some in the westerly direction (navigational format),
the longitude of the observer M is the same in both systems.

= E and W = ; W = E:
> 0 , Easterly; < 0 , Westerly
D
l (or or ) is de…ned as l = with and in customary mathematical
format for measuring angles, that is, measured positively in an easterly direction.
Angles in celestial navigation are traditionally measured in navigational format, that
is, positively in a westerly direction.
Most angles (with one exception) tabulated or computed in celestial navigation
are positive. If a computed angle is negative, it is changed to a positive angle by

4
adding to it 360 . Moreover if an angle is greater than 360 , we subtract 360 from
it.

Let LHA(P ) = 360 , the local hour angle of P

GHA = 360 ) = 360 GHA

Then
LHA = 360 ( )
LHA = + (360 )= + GHA

l is the angle by which the celestial body P is east of the observer M .

LHA is the angle by which celestial body P is west of observer M .

We may then write

LHA(P ) = GHA (P ) + (2)

For example, see Figure 3 below.

Figure 3: The Equality of 360 and LHA.

If the three points on the globe, the North Pole, M and P are connected by
great circles, there are two possible navigational (spherical) triangles. In celestial
navigation, we are interested in the smaller of these, the spherical triangle which has
the smaller angle between the meridian of the celestial body and the meridian of the

5
observer. However, the equations in this paper derived via rotation matrices apply
to any spherical triangle.
The longitude di¤erence LHA is frequently supplemented by the measure t " [0; 180 ),
the meridian angle, the smaller of the two angles between the meridian of the observer
M and the meridian of the geographical position of the celestial body observed [10].
It is measured east or west, t = tE or tW , depending upon its value.
If LHA 180 , then tW = LHA, celestial body P west of observer’s meridian.
If LHA > 180 , then tE = 360 LHA, celestial body P east of observer’s meridian.
Both of these meridian angles are positive.
In this paper all declinations north of the equator are positive; those south of
the equator are negative. Until relatively recently, before modern calculators and
computers were available, arguments of trigonometric functions were tabulated for
angles in the …rst trigonometric quadrant, that is, 0 to 90 . If any sign changes
of the trigonometric functions of angles used in the navigational calculations were
necessary for angles residing in any quadrant other than the …rst, rules were used to
assign these signs.

1.1 Derivation of The Navigation Equations

In the derivation of the equations of spherical trigonometry used in celestial naviga-
tion, there are 3 rotations of coordinates to be performed.
I typically use the symbols (x; y; z) to refer to coordinates of a vector R in a
coordinate system S and (x0 ; y 0 ; z 0 ) to refer to coordinates of the same vector R in
the rotated coordinate system S 0 (rotated relative to system S). Frequently, system
S is referred to as the "laboratory system" with (x; y; z) as the "space axes" and
system S 0 as the "body axis system" with (x0 ; y 0 ; z 0 ) as the "body axes". Here, the
laboratory system is the celestial equator (equinoctial) system and the "body axis
system" is the horizon system.
I know A and A : I want to determine B and B , or conversely.
We proceed as follows:

1.1.1 Sequence of Three Rotations to be Performed

For the purpose of this section, we will undertake a temporary reassignment of symbols
for the Cartesian coordinates involved.
Step 1. The Cartesian coordinates in the equinoctial system of celestial body P
are (x0 ; y0 ; z0 ). The …rst rotation is around the z0 axis by angle = E (M ), the east
longitude of the observer M. The new coordinates of P are (x1 ; y1 ; z1 ). (At the end
of these three coordinate rotations, we will relabel (x0 ; y0 ; z0 ) as (x; y; z)).
Step 2. The second rotation is around the y1 axis by angle , the colatitude of
the observer M. The new coordinates of P are (x2 ; y2 ; z2 ).

6
 Step 3. The third rotation is around the z2 axis by angle = 180 , so that the
new x-axis points in the northerly direction. The new coordinates of P are (x3 ; y3 ; z3 ).

Figure 4: After Two Rotations.

Now, reassign (x0 ; y0 ; z0 ) (x; y; z) and (x3 ; y3 ; z3 ) (x0 ; y 0 ; z 0 ).

Proceeding in this way as we have done before, we write
0 1 0 1
x sin A cos A
Suppose that @ y A = R @ sin A sin A A are the components of a vector R in
z cos A
the celestial equator (equinoctial system), and that this coordinate system is copied

and then rotated via G( ; ; ) = Z( ) Y ( ) Z( ), which we call the gyro rotation

matrix.

0 10 10 1
cos sin 0 cos 0 sin cos sin 0
G( ; ; ) = @ sin cos 0 A @ 0 1 0 A @ sin cos 0 A
0 0 1 sin 0 cos 0 0 1

0 1
cos cos cos sin sin cos sin + cos cos sin sin cos
= @ sin cos cos cos sin cos cos cos sin sin sin sin A (3)
sin cos sin sin cos

7
 After the coordinate system has been rotated via Z( ) and Y ( ) , the new (car-
ried) x-axis will be pointing away from the original z-axis, that is, it will be pointing
in a south direction along the meridian to which it is tangent. But we require that
the x-axis point in a northerly direction along the meridian, because North is the di-
rection from which co-azimuth is measured. For this to occur, the coordinate system
must be rotated by around the latest z-axis, that is, we must have = 180 . (This
was described above).
0 1
cos cos cos sin sin
G( ; ; ) = Z( ) Y ( ) Z( ) = @ sin cos 0 A (4)
cos sin sin sin cos
Figure 5 portrays the relevant lines and angles with which we are concerned. Keep
in mind that the axes x0 and y 0 lie in a di¤erent plane than axes x and y.

Figure 5: Position Vectors of the GP and the Observer’s Position M.

8
 The Euler rotation matrix rather than the Gyro rotation matrix can be used for
these derivations with ultimately the same results.
Diagrams or …gures constructed for the study of spherical trigonometry portray
relevant lines, angles and great circles. These enable us to visualize the mutual
geometrical relationships of these lines, angles and great circles, and to subsequently
declare these relationships algebraically. Without such …gures, it would be di¢ cult
to accomplish this task. Moreover, these …gures, which represent three-dimensional
entities, are produced on a two-dimensional sheet of paper as perspective drawings.
We do not have three-dimensional (e.g., holographic) drawings, and visualization
of perspective on a two-dimensional surface can be di¢ cult if care is not taken in
their creation. Furthermore, the …gures can become cluttered if we attach all of
the relevant lines and symbols, detracting from their visualization. For example,
the angle of intersection of two great circles is measured by the angle between their
tangents at the point of intersection. Conventionally however, we usually express this
angle as between the circular arcs themselves as illustrated below.

Figure 6: Arcs and Tangents.

In Figure 7A and 7B below we observe that the vertices of the spherical triangle
are each connected to the origin O by equal radii R. The vertices are N P , M and P .

9
 Figure 7A: Navigational Triangle, P East of M.

Figure 7B: Navigational Triangle, M East of P.

The intersection of the three great circles spanned by the central angles ; ;

10
and the surface of the earth form a trihedron. If we include the chords and/or arcs
between M; P and N P , we have a tetrahedron.
(e.g., the …gure above created from line segments OM ; OP ; O N P ; M P ; N P M ;
N P P and the corresponding great circle arcs). These and the spherical triangle
M-P-NP possess threefold symmetry.
The angles ( ; ) of Figure 9A or ( ; LHA) of Figure 9B each subtend the
circular arc of length R :
The angles ( ; Z) of Figure 9A or ( ; A) of Figure 9B each subtend the circular
arc of length R :
The angles ( ; ) of Figure 9A or ( ; B) of Figure 9B each subtend the circular
arc of length R :
Equations derived from the analysis of the tetrahedron alone are the same for the
three angle pairs except for their interior angle arguments. As will be shown shortly,
because of the symmetry inherent in the tetrahedron, we may permute the symbols
in equations 4, 5, 6. These equations are known as the spherical trigonometric sine
and cosine equations. Derivations of these appear in the Appendix.

However, when the tetrahedron is embedded in the Cartesian coordinate system

S overlaid by spherical coordinates (using the results of the coordinate rotations):
1. The same cosine equations continue to be threefold symmetrical, except now
their arguments also include exterior angles. These angles are not the same as the
interior angles, but are arithmetically related to them.
2. Three of the sine equations are symmetrical.
3. Additional equations are generated via the coordinate rotation process. These
provide information to uniquely justify the trigonometric quadrants and are not usu-
ally derived via the "classical" method appearing in the Appendix.

If the radii and chords of the tetrahedron were characterized by (overlaid with)
direction cosines rather than by spherical coordinates, there would be complete three-
fold symmetry, because direction cosine angles are all measured in the same manner.
However, the two spherical coordinate angles (colatitude and longitude) are not mea-
sured in a similar manner to one-another. Longitude for both M and P are measured
in the same plane, whereas the co-declination and co-latitude are each measured in
di¤erent planes.

1.2 A Perspective on the Measurement Angles

In Figure 8, the surface of the earth in the neighborhood of the observer’s position
M is represented by the surface of the rotor of a mechanical gyroscope. This is a
supplementary …gure introduced here merely to provide, in this author’s opinion, a
better visualization of the relevant angles. In the neighborhood of M, the surface of
the earth may, for computational purposes, be regarded as being ‡at.

11
 Figure 8: Celestial Equator and Horizon Systems of Coordinates.

Referring to Figure 5, suppose that we replace symbols,

, = (unchanged), A , B , A , B A

The components of vector R in the original Cartesian coordinate system (the

celestial equator system) are

0 1 0 1 0 1
x sin A cos A sin cos
@ y A = R @ sin A sin A
A = R @ sin sin A (5)
z cos A cos
The components of vector R in the new rotated Cartesian coordinate system (the
horizon system) are

12
 0 1 0 1 0 1
x0 sin B cos B sin cos A
@ y 0 A = R @ sin B sin B
A = R @ sin sin A A (6)
z0 cos B cos

0 1 0 1
x0 x
@ y 0 A = G( ; ; ) @ y A (7)
z0 z
where G( ; ; ) = G( ; ; ). That is,

0 1 0 10 1
sin cos A cos cos cos sin sin sin cos
@ sin sin A A = @ sin cos 0 A @ sin sin A (8)
cos cos sin sin sin cos cos
0 1
cos sin cos cos cos sin cos sin sin sin
=@ cos sin sin cos sin sin A
cos cos + cos cos sin sin + sin sin sin sin
0 1
cos sin cos sin (cos cos + sin sin )
=@ sin (sin cos cos sin ) A
cos cos + sin sin (cos cos + sin sin )
0 1 0 1
sin cos AB cos sin cos sin cos( )
That is, @ sin sin A A = @ sin sin( ) A
cos cos cos + sin sin cos( )
0 1 0 1
sin cos A cos sin cos sin cos l
@ sin sin A A = @ sin sin l A , where l = =
cos cos cos + sin sin cos l

Displayed below are the equations of spherical trigonometry using di¤erent com-
binations of (alias) symbols, (e.g., l = = ).
Angles measured in standard mathematical format, that is, easterly:

sin cos B = cos A sin cos sin A cos (9.1)

sin sin B = sin A sin (9.2)

cos = cos A cos + sin A sin cos (9.3)

where = A = (alias l or )

13
 Displayed below are the equations of spherical trigonometry using di¤erent com-
binations of (alias) symbols:

sin cos A = cos sin cos sin cos l (10.1)

sin sin A = sin sin l (10.2)

cos = cos cos + sin sin cos l (10.3)

where l = = =

If the third equation 10.3 is solved for cos l and substituted into equation 10.1, we
obtain
cos cos cos
cos A = (11)
sin sin

sin sin l
sin A = (12.1)
sin

cos cos cos

cos A = (12.2)
sin sin

cos = cos cos + sin sin cos l (12.3)

where l = =

Angles measured in navigational format with LHA = 360 l, that is, westerly:

cos h cos Z = sin d cos L cos d sin L cos LHA (13.1)

cos h sin Z = cos d sin LHA (13.2)

sin h = sin d sin L + cos d cos L cos LHA (13.3)

where LHA = 360 l

If the third equation 13.3 is solved for cos LHA and substituted into equation
13.1, we obtain
sin d sin L sin h
cos Z = (14)
cos L cos h

14
 hc is computed altitude, d (or alias ) is declination of celestial body P, and L
is latitude of observer M. Let subscript "a" indicate "assumed", and subscript "c"
indicate "computed".

Angles measured in navigational format, that is, westerly:

cos d sin LHA

sin Z = (15.1)
cos h
sin d sin L sin h
cos Z = (15.2)
cos L cos h

sin h = sin d sin L + cos d cos L cos LHA (15.3)

where LHA = GHA +

In the practice of celestial navigation, we usually use equations 15.1,

15.2 and 15.3. And, of course, equation (15.3) must be evaluated …rst for subsequent
substitution in to (15.1) and (15.2). L is the assumed latitude of observer M, is
the assumed longitude of observer M, with westerly observer longitudes replaced by
their negatives. h is computed altitude of celestial body P. A subscript "c" might
typically be appended to these symbols to represent "computed".

1.3 More Spherical Trigonometry Equations

Figures 9A, 9B, 10A, 10B below provide re…nements to Figures 7A, 7B presented
earlier. In these …gures we notice a threefold symmetry.
The internal angles ; ; subtend circular arcs. The internal angles ; Z; of
Figure 9A or LHA; A; B of Figure 9B are angles on the surface of the celestial sphere
between tangent lines of intersecting circular arcs.

15
 Figure 9A:

Figure 9B:

Celestial Body Coordinates Revisited

From previous derivations, we have:

Coordinates of celestial body P (of GP) relative to Greenwich G in system S,

16
0 1 0 1
x sin cos
@ y A = R @ sin sin A (5, repeated);
z cos

0 Coordinates
1 0 of celestial
1 body P (of GP) relative to observer M in system S 0 ,
x0 sin cos A
@ y 0 A = R @ sin sin A A (6, repeated);
z0 cos
0 0 1 0 1
x x
@ y 0 A = G( ; ; ) @ y A
z0 "
Coordinates
z
" of M "
P osition of P P osition of P
in S 0 system in S system
(W hat M sees) (W hat G sees)

0 1 0 1
sin cos A sin cos
that is, @ sin sin A A = G( ; ; ) @ sin sin A
cos cos
where 0 1
cos cos cos sin sin
G( ; ; ) = @ sin cos 0 A
sin cos sin sin cos
0 1 0 10 1
sin cos A cos cos cos sin sin sin cos
@ sin sin A A = @ sin cos 0 A @ sin sin A
cos cos sin sin sin cos cos

We may now consider the inverse relationship,

0 1 0 1 1
0
x x0 x0
@ y A = G 1( ; ; ) @ y0 A GT ( ; ; ) @ y 0 A
z z0 z0
0 1 0 10 1
sin cos cos cos sin cos sin sin cos A
@ sin sin A = @ cos sin cos sin sin A @ sin sin A A
cos sin 0 cos cos
0 1
sin sin sin A + cos (sin cos cos sin cos A)
=@ cos sin sin A + sin (sin cos cos sin cos A) A )
cos cos + sin sin cos A

sin cos = sin sin sin A + cos (sin cos cos sin cos A) (16.1)
sin sin = cos sin sin A + sin (sin cos cos sin cos A) (16.2)
cos = cos cos + sin sin cos A (16.3)

17
 (Eq:16:1) cos + (Eq:16:2) sin )

sin cos l = sin cos cos sin cos A (17)

where l =

Observer Coordinates Revisited

1 0 0 1
u sin cos
Coordinates of observer M in system S, @ v A @
= R sin sin A (18)
w cos
0 0 1 0 1
u sin cos
0 @ 0 A @ A (19)
Coordinates of observer M in system S , v = R sin sin
w0 cos
0 0 1 0 1
u u
@ v 0 A = G( ; ; ) @ v A (20)
0 "
w Coordinates
w
" of P "
P osition of M P osition of M
in S 0 system in S system
(W hat P sees) (W hat G sees)

0 1
cos cos cos sin sin
G( ; ; ) = @ sin cos 0 A (21)
cos sin sin sin cos
0 1 0 10 1
sin cos cos cos cos sin sin sin cos
@ sin sin A=@ sin cos 0 A @ sin sin A (22)
cos cos sin sin sin cos cos

0 1
sin cos
cos sin cos( )
=@ sin( ) sin A , that is,
cos cos + cos( ) sin sin
0 1 0 1
sin cos sin cos cos sin cos l
@ sin sin A = @ sin l sin A
cos cos cos + sin sin cos l

sin cos = sin cos cos sin cos l (23.1)

sin sin = sin sin l (23.2)
cos = cos cos + sin sin cos l (23.3)

18
With the exception of a di¤erent variable (requiring a minus sign) in equation (10.2),
equations (23.2) and (10.2) are of the same form. With the substitution of A =
360 Z into equation 10.2 we may restore the symmetry.

sin sin Z = sin sin l (24)

Equation (23.3) is identical to (10.3).

From Equations 10.2 and 23.2, we observe that

sin l sin sin Z

= = (25)
sin sin sin

The three equations 25 are known as the Spherical Law of Sines. In Figure 10A, we
see that these are interior angles of the navigational tetrahedron.
0 1 0 0 1
u u
Again, we may consider the inverse relationship, @ v A = G ( ; ; ) v 0 A.
T @
w w0

0 1
cos cos sin cos sin
G 1 ( ; ; ) = GT ( ; ; ) = @ cos sin cos sin sin A
sin 0 cos
0 1 0 10 1
sin cos cos cos sin cos sin sin cos
@ sin sin A=@ cos sin cos sin sin A @ sin sin A
cos sin 0 cos cos
0 1 0 1
sin cos sin sin sin + cos (sin cos cos sin cos )
@ sin sin A = @ sin sin cos sin (cos sin + cos sin cos ) A
cos cos cos + sin sin cos
That is,

sin cos = sin sin sin + cos (sin cos cos sin cos ) (26.1)
sin sin = sin sin cos sin (cos sin + cos sin cos ) (26.2)
cos = cos cos + sin sin cos (26.3)

Equations 10.3, 16.3 and 26.3 are known as the Spherical Law of Cosines and are,
for convenience, restated below.

cos = cos cos + sin sin cos l (10.3, repeated)

cos = cos cos + sin sin cos A (16.3, repeated)
cos = cos cos + sin sin cos (26.3, repeated)

19
Note: cos Z cos A, so that (16.3) may used with that substitution if required.
Elsewhere, sin Z = sin A may be used.
A traditional derivation of the spherical law of sines and of cosines is provided in
the Appendix.

1.4 A Second Spherical Law of Cosines

We have already derived equations (10.3), (16.3) and (26.3), the three instances
of the spherical law of cosines. These equations are repeated below for convenience.

cos = cos cos + sin sin cos l (10.3, repeated)

cos = cos cos + sin sin cos A (16.3, repeated)
cos = cos cos + sin sin cos (26.3, repeated)

These equations express the cosines of the interior angles ; and as functions of the
other interior angles and the one surface angle l; Z; or , repectively, corresponding
to the interior angle on the left hand side of these equations. There exist three
complementary or converse equations of rather similar but not identical form, with
interior angles and surface angles interchanged. These equations are provided below,
followed by their derivations from equations (10.3), (16.3) and (26.3). These converse
equations express what we refer to as the second spherical law of cosines.

cos l = cos Z cos + sin Z sin cos (27.1)

cos Z = cos cos l + sin sin l cos (27.2)
cos = cos l cos Z + sin l sin Z cos (27.3)

1.4.1 Some Auxiliary Equations

We …rst derive a set of equations from (10.3), (16.3) and (26.3), which equations will
subsequently be used to derive (27.1), (27.2) and (27.3).
Derivation 1.
Substitute (3) into (1) to eliminate cos :

cos = cos (cos cos + sin sin cos ) + sin sin cos l

cos = cos cos2 + sin sin cos cos + sin sin cos l
cos 1 cos2 = sin sin cos cos + sin sin cos l
cos sin2 = sin sin cos cos + sin sin cos l
cos sin = sin cos cos + sin cos l

sin sin
cos = cos cos + cos l
sin sin

20
But
sin l sin Z sin sin sin Z sin sin
= = (4) ) = and =
sin sin sin sin sin l sin sin l
Making these substitutions,
sin Z sin
cos = cos cos + cos l
sin l sin l
cos sin Z = cos sin l cos + cos l sin (28.1)

Derivation 2.
Substitute (3) into (2) to eliminate cos :

cos = (cos cos + sin sin cos ) cos + sin sin cos Z
cos = cos2 cos + sin cos sin cos + sin sin cos Z
cos 1 cos2 = sin cos sin cos + sin sin cos Z

cos sin2 = sin cos sin cos + sin sin cos Z

cos sin = cos sin cos + sin cos Z
sin sin
cos = cos cos + cos Z
sin sin
But
sin l sin Z sin sin sin l sin sin
= = (4) ) = and =
sin sin sin sin sin Z sin sin Z

Making these substitutions,

sin l sin
cos = cos cos + cos Z
sin Z sin Z
cos sin l = cos sin Z cos + sin cos Z 2nd eq. (28.2)

Derivation 3.
Substitute (2) into (3) to eliminate cos :

cos = cos (cos cos + sin sin cos Z) + sin sin cos
cos = cos cos2 + sin sin cos cos Z + sin sin cos
cos 1 cos2 = sin sin cos cos Z + sin sin cos

21
 cos sin2 = sin sin cos cos Z + sin sin cos
cos sin = sin cos cos Z + sin cos
sin sin
cos = cos cos Z + cos
sin sin
But
sin l sin Z sin sin sin l sin sin Z
= = (4) ) = and =
sin sin sin sin sin sin sin

Making these substitutions,

sin l sin Z
cos = cos cos Z + cos
sin sin
1

cos sin l = cos cos Z sin + sin Z cos (28.3)

Derivation 4.
Substitute (2) into (1) to eliminate cos :

cos = (cos cos + sin sin cos Z) cos + sin sin cos l

cos = cos2 cos + sin cos sin cos Z + sin sin cos l
cos 1 cos2 = sin cos sin cos Z + sin sin cos l
cos sin2 = sin cos sin cos Z + sin sin cos l
cos sin = cos sin cos Z + sin cos l
sin sin
cos = cos cos Z + cos l
sin sin
But
sin l sin Z sin sin sin sin sin Z
= = (4) ) = and =
sin sin sin sin sin l sin sin l
Making these substitutions,

sin sin Z
cos = cos cos Z + cos l
sin l sin l

cos sin = cos sin l cos Z + sin Z cos l (28.4)

22
 Derivation 5.
Substitute (1) into (2) to eliminate cos :

cos = cos (cos cos + sin sin cos l) + sin sin cos Z

cos = cos cos2 + sin sin cos cos l + sin sin cos Z
cos 1 cos2 = sin sin cos cos l + sin sin cos Z
cos sin2 = sin sin cos cos l + sin sin cos Z
cos sin = sin cos cos l + sin cos Z
sin sin
cos = cos cos l + cos Z
sin sin
But
sin l sin Z sin sin sin sin sin Z
= = (4) ) = and =
sin sin sin sin sin l sin sin l
Making these substitutions,

sin sin Z
cos = cos cos l + cos Z
sin l sin l

cos sin = sin Z cos l cos + sin l cos Z (28.5)

Derivation 6.
Substitute (1) into (3) to eliminate cos :

cos = (cos cos + sin sin cos l) cos + sin sin cos
cos = cos2 cos + sin cos sin cos l + sin sin cos

cos 1 cos2 = sin cos sin cos l + sin sin cos

cos sin2 = sin cos sin cos l + sin sin cos
cos sin = cos sin cos l + sin cos
sin sin
cos = cos cos l + cos
sin sin
But
sin l sin Z sin sin sin Z sin sin
= = (4) ) = and =
sin sin sin sin sin l sin sin l

23
Making these substitutions,
sin Z sin
cos = cos cos l + cos
sin l sin l
cos sin Z = cos sin cos l + sin l cos (28.6)
Summary of Auxiliary Equations:

cos sin Z = cos sin l cos + cos l sin (28.1)

cos sin l = cos sin Z cos + sin cos Z (28.2)

cos sin l = cos cos Z sin + sin Z cos (28.3)

cos sin = cos sin l cos Z + sin Z cos l (28.4)

cos sin = sin Z cos l cos + sin l cos Z (28.5)

cos sin Z = cos sin cos l + sin l cos (28.6)

1.4.2 Combining the Auxiliary Equations

First Instance of the Second Cosine Law:

cos sin Z = cos sin l cos + sin cos l (28.1)

cos sin l = cos sin Z cos + sin cos Z (28.2)
Substitute (28.2) into (28.1) to eliminate cos sin l:

cos sin Z = (cos sin Z cos + sin cos Z) cos + cos l sin
cos sin Z = cos sin Z cos2 + sin cos cos Z + cos l sin
cos sin Z 1 cos2 = sin cos cos Z + cos l sin
cos sin Z sin2 = sin cos cos Z + cos l sin
cos sin Z sin = cos cos Z + cos l
) cos l = cos Z cos + sin Z sin cos (27.1, repeated)

Second Instance of the Second Cosine Law:

cos sin = sin Z cos l cos + sin l cos Z (28.5)

24
 cos sin Z = cos sin cos l + sin l cos (28.6)
Substitute (28.6) into (28.5) to eliminate cos sin Z:

cos sin = (cos sin cos l + sin l cos ) cos l + sin l cos Z

cos sin = cos sin cos2 l + sin l cos l cos + sin l cos Z
cos sin 1 cos2 l = sin l cos l cos + sin l cos Z
cos sin sin2 l = sin l cos l cos + sin l cos Z

cos sin sin l = cos l cos + cos Z

) cos Z = cos l cos + sin sin l cos (27.2, repeated)

Third Instance of the Second Cosine Law:

cos sin = sin Z cos l cos + sin l cos Z (28.5)

cos sin Z = cos sin cos l + sin l cos (28.6)

Substitute (28.5) into (28.6) to eliminate cos sin :

cos sin Z = (sin Z cos l cos + sin l cos Z) cos l + sin l cos
cos sin Z = (sin Z cos l cos + sin l cos Z) cos l + sin l cos
cos sin Z = sin Z cos2 l cos + sin l cos l cos Z + sin l cos
cos sin Z 1 cos2 l = sin l cos l cos Z + sin l cos
cos sin Z sin2 l = sin l cos l cos Z + sin l cos
cos sin Z sin l = cos l cos Z + cos
) cos = cos l cos Z + sin l sin Z cos (27.3, repeated)

25
2 Right Triangles and Short-Method Tables
For three points on the surface of the earth, N P , M and P , there are two possible
polar spherical triangles, the smaller (minor) local triangle and the larger (major)
remote triangle on the other side of the earth. The coaltitude spans the smaller
spherical triangle, and it is this spherical triangle with which we are concerned. The
major spherical triangle, spanned by 360 , is of no interest to us here.
The spherical triangles of Figures 7A and 7B can be cut with great circles into
two adjacent right spherical triangles via an auxiliary great circle. An advantage of
doing so is that the equations describing them are simpli…ed. The sine of 90 degrees
is one and the cosine of 90 is zero, rendering one or more terms in the equations
constant, 0 or 1.
Several relatively short mathematical tables have been compiled using these equa-
tions to "solve the navigational triangle", that is, to compute the altitude and azimuth
of a celestial body based upon its LHA and the observer’s latitude L.

Two di¤erent methods are employed to accomplish this based upon an auxiliary
great circle con…gured in one of two ways. In the past, a number of investigators
have derived and published their short-method trigonometric derivations and tables,
but all are somewhat similar to one-another.

Ogura’s Method: The new great circle passes through the observer’s position
M and intersects the declination great circle at a right angle. Reference: Figures 10A
and 10B. This method is employed in the short-method navigational tables of Sintiki
Ogura (1884-1937), the Line of Position Book [11] of P.V.H Weems, H.O. 208 of
Dreisonstock, among others.

Ageton’s Method: The new great circle passes through the geographical posi-
tion P (or GP ) of the celestial body and intersects the colatitude great circle at a
right angle. Reference: Figures 11A and 11B.
This method is employed in H.O 211 by Arthur Ageton (1931) and a modi…ed
H.O 211 by Allan Bayless (1980).
"H.O." is an abbreviation for "Hydrographic O¢ ce", the name of the American
government agency that formerly dealt with navigational charts.

In Figures 10A, 10B, 11A and 11B, note the arcs of great circles between M
and P (Ogura’s Method) and between P and M 0 (Ageton’s Method), (
0
and
respectively). As described above, these great circles cut the navigational triangles
into two adjacent right-angled spherical triangles. The two companion right-angled
spherical triangles each have more simple algebraic and trigonometric forms than the
original triangles, and are used in compiling the "short-method" mathematical tables
for celestial navigation.

26
 The original polar spherical triangle and the upper right-angled polar spherical
triangle can be analyzed with the original equations (15.1, 15.2, 15.3). Both of these
spherical triangles are directly described via spherical coordinates. The upper right-
angled polar spherical triangle with vertices N P; M; P 0 is sometimes referred to as
the Time triangle [6], and the lower right-angled non-polar spherical triangle with
vertices P 0 ; M; P is sometimes referred to as the Altitude triangle [6]. The altitude
triangle cannot be directly described via spherical coordinates in the same manner as
the other two triangles, because it does not have a vertex at N P .
For economy of space in the …gures, the single letter symbols M and P are used.
M represents the observer’s position, usually designated as AP (assumed position),
and P represents the geographical position of the celestial body, usually designated
as GP . The symbols l and are aliases of each other.

In Figures 10A and 10B, the expressions "P East of M" and "P West of M" simply
indicate that these …gures represent the extreme con…gurations of the spherical trian-
gle(s) for small and large l , i.e., for large and small GHA. The same trigonometric
equations are used to "solve" these spherical triangles.

The domain of the coaltitude , as a standard spherical coordinate, is [0; 180 ].

However, in marine celestial navigation, since = 90 h, is further restricted to
[0; 90 ]. Altitudes h below the horizon are not typically measured by the marine
sextant. Hence, if we know sin , we can uniquely determine , an angle in the …rst
trigonometric quadrant.

27
 Figure 10A: Spherical Triangle, P East of M.

Figure 10B. Spherical Triangle 1, P West of M.

28
2.1 Splitting the Codeclination, Ogura’s Method
For the upper spherical triangle, we use the spherical sine formulas.
sin l sin sin Z
= = (25, repeated)
sin sin sin
Domains of the angles:

D( ) = [0; 180 ] ) ( ) : sin 0

D( ) = [0; 180 ] ) ( ) : sin 0

However, D( ) = D(l) = [0; 360 ], so sin and sin l may be positive or negative.

But for the ratios

sin sin sin 0 sin
= 0 and = 0
0
sin l sin sin l sin

) sign(sin ) = sign(sin l):

0 0 0 0
= 90 ) sin = 1 and = 270 ) sin = 1:

0 sin l
sin = sin = sin jsin lj 0 (29)
sin 0
0 0
sin sin = sin sin l and sin sin = sin sin l
0 0
sin sin = sin sin
0
sin sin
= 0
sin sin
Observe in Figure 10A that as l decreases with M moving to the right, increases
and P 0 moves down the codeclination great circle and 0 increases. When l = 0 (i.e.,
360 ), = 180 , M and P lie on the same meridian. As M moves further to the
right (east) of P and l decreases from 360 , decreases, the point P 0 moves up the
codeclination great circle, 0 decreases until …nally 0 = 270 .
As M moves from its original position to the left (westward) from P , l increases
and 0 decreases. When l = 180 , P 0 moves up to N P . The angle 0 is always
equal to 90 or 270 .
sin
sign 0 = sign . Therefore, 0 and sign (sin 0 ) = sign (sin ).
sin 0
Derivation.

sin cos = sin cos cos sin cos l (23.1, repeated)

29
 0 0 0 0
sin cos = sin cos cos sin cos l
0 0 0
= 90 or = 270 ) cos = 0.

) sin 0 cos = cos 0 sin cos l (30)

) tan 0 = tan cos l
0
is the codeclination of point P 0 and its domain is [0; 180 ]. If tan 0 0,
then 0 lies in the 1st trigonometric quadrant; if tan 0 < 0, then 0 lies in the 2nd
0
trigonometric quadrant. Thus, = arctan (tan 0 ) uniquely determines the value
0
of the angle .

Derivation.

cos = cos cos + sin sin cos (26.3, repeated)

0 0 0 0 0
cos = cos cos + sin sin cos
) cos = cos 0
cos 0

0 cos
cos = 0
(31)
cos
0 0
Substituting cos = cos cos into equation (30), we have...

0 0
sin cos = sin cos l (32)

Derivation.

cos = cos cos + sin sin cos l (10.3, repeated)

0 0 0 0
From (31), cos = cos cos and from (32), sin cos = sin cos l. Making
these replacements,
0
cos = cos (cos cos 0 ) + sin (sin 0
cos 0 )
0 0 0
cos = cos (cos cos + sin sin )
) cos = cos 0
cos ( 0
) (33)
0 cos 0
Since the cosine function cos ( ) = 0
and cos ( ) < 1, so
cos
cos
< 1 and cos < cos 0 ) 0
< .
cos 0
Derivation.

30
 Recall
sin cos = sin sin sin + cos (sin cos cos sin cos ) (26.1, repeated)
sin sin = sin sin cos sin (cos sin + cos sin cos ) (26.2, repeated)
cos = cos cos + sin sin cos (26.3, repeated)
0 0 0
sin = 1, depending upon with cos = 0. Therefore,

sin cos = sin 0 sin 0

sin 0 + cos 0
sin 0
cos 0
0
sin sin = sin 0 sin 0
cos 0 sin 0
cos 0
sin
cos = cos 0 cos 0

Re-arrange terms in second equation:

sin cos = sin 0 sin 0 sin 0 + cos 0 sin 0 cos 0

sin sin = cos 0 sin 0 sin 0 + sin 0 sin 0 cos 0

cos = cos 0 cos 0

Multiplying the 1st equation by sin and the 2nd equation by cos ;

sin cos sin = sin 0 sin 0 sin 0 sin + cos 0 sin 0 cos 0 sin
sin sin cos = cos 0 sin 0 sin 0 cos + sin 0 sin 0 cos 0 cos
cos = cos 0 cos 0

Subtract the 2nd equation from the 1st equation:

0 0 0 0 0 0 0 0
0 = sin sin sin sin + cos sin sin cos + cos sin cos 0 sin
0 0
sin sin cos 0 cos
0 0 0 0 0 0 0 0
0 = sin sin (cos cos + sin sin ) sin cos (sin cos cos sin )
0 0 0 0 0 0
sin sin cos ( ) = sin cos sin ( )
0 0 0 0
tan sin = sin tan ( )
0
But = and l = , so
0 0 0
tan sin = sin tan l (34 )

31
2.1.1 The Line of Position Book
In the Line of Position Book (LPB), Weems uses the methods of Sintiki Ogura to split
the spherical navigational triangle into two right-angled spherical triangles [11]. The
right angles occur on the declination great circle. Dreisonstock used a similar method
for his tables, H.O. 208. The LPB splits the calculations into two parts, each part
using fairly simple equations to solve the spherical triangle. Since the Line of Position
Book tables are tabulated only for " (0; 180 ), the meridian angle t (rather than
LHA) and observer’s latitude L (DR or AP ) are input into the LPB tables. Upon
completing the two part calculations, the …nal outputs are the computed altitude hc
and azimuth Z of the celestial body. However, in our derivations, we will use the
more general LHA rather than meridian angle t.
Ogura’s method and similar methods were developed before the advent of modern
digital computers. In legacy practice of spherical trigonometry and celestial naviga-
tion, cosecants and secants were usually employed instead of sines and cosines, because
these functions always possess values greater than or equal to one. Furthermore, the
105 factor was used to generate large integers. It is peculiar that in the LPB, Weems
and Lee omit mention of the 105 factor in their explanation. Logarithms were then
used to evaluate products or ratios of these positive integers; this practice, done with
paper and pencil, made the computational process possible [11].

First Part Calculation. Entering with LHA and L, the output from the …rst
part calculation consists of two numbers, A and K, where

A = 105 log sec 0

= 105 log sec (90 h0 ) = 105 log csc h0 (35.1)
K = d0 = 90 0
(35.2)
in which we have written the symbols A and K in script (calligraphic) form to
distinguish the …rst symbol from the symbol A which is already used to represent the
0
coazimuth of P . and 0 are the coaltitude and codeclination that the celestial
body would have if its altitude great circle and declination great circle intersected at
90 as shown in Figure 10A, that is, if it were located at point P 0 in that …gure.
Hence, (LHA; L) ) (A; K) ) (h0 ; d0 ), that is, (l; ) ) ( 0 ; 0 ). The numbers A
and K are obtained from Table A of the LPB.

Compute 0 .
0
The LPB uses (29) with L = 90 to evaluate .
0
sin = sin jsin lj (29, repeated)

In the First Part Calculation in the Line of Position Book, the angle 0 is calculated
simply by taking the arcsine of sin 0 . This is su¢ cient to uniquely evaluate the angle
0
, because in marine navigation 0 is always an angle in the …rst trigonometric
quadrant, that is, less than or equal to 90 .

32
For aerial or space navigation, in which we might measure negative sextant angles
(i.e., below the local horizontal plane), the coaltitudes would be angles in the second
trigonometric quadrant. Keep in mind that the angles and 0 are normal spherical
coordinates just like and whose mathematical domains are [0; 180 ]. In that case,
equation 28 alone would be insu¢ cient to uniquely determine 0 . We would require
quadrant justi…cation via, e.g., the cosine or the tangent function. This situation will
be addressed shortly.

Compute d0 .
0
Having just evaluated , equation 31 is now used to determine d0 .

0 cos
cos = 0
(31, repeated)
cos
d0 = 90 0

End of …rst part calculations.

Second Part Calculation.

De…nitions:
D
K d = d0 d = (90 0
) (90 )= 0
(35.3)

If the declinations d and d0 are South declinations, they are negative numbers.

D 1
B = 105 log sec (K d) = 105 log sec ( 0
) = 105 log 0)
(35.4)
cos (

Again, a script (calligraphic) font is used for this quantity to distinguish it from the
angle B.
K d and B are obtained from Table B of the LPB. A and B are then added
together.
A + B = 105 log sec 0 + 105 log sec ( 0
)
= 105 log [sec 0
sec ( 0
)]
But
0 0
cos = cos cos ( ) (33, repeated)

33
so
0 0
sec = sec sec ( )
and
A + B = 105 log sec

The sum A and B is then entered into Table B of the LPB to obtain the computed
altitude h. The degrees of altitude appear on the bottom of the page and the minutes
of altitude in the right margin.
.

Azimuth Calculation. In the Line of Position Book, the azimuth Z of P is

sin l sin Z
obtained from "Rust’s Diagram". This diagram is based upon (25) =
sin sin
sin sin l
) sin Z = , except that Weems uses meridian angle t (i.e., HA) instead
sin
of l.

End of LPB calculations (Weems).

2.1.2 Numerical Example

Suppose that L = 35 N ) = 55 , LHA = 48 ) l = 312 .
d = 20 S = 20 ) = 110 .
sin 0 = sin jsin lj = sin 55 jsin 312 j = (0:8195) j 0:74314j = 0:6088
0
= arcsin (sin 0 ) = 37:499 or 142:501 . In marine celestial navigation we are
only interested in the value in the …rst trigonometric quadrant, so 0 = 37:50 . The
LPB goes no further in trigonometric quadrant determination of angle 0 .
cos cos 55
From (31), cos 0 = 0
= = 0:7230, and 0
= arcos(0:7230) =
cos cos 37:50
43:70 .
0
= 110 43:70 = 66:30 .
From (33), cos = cos 0 cos ( 0
) = cos 37:50 cos (66:30 ) = 0:3189, and
= arcos(0:3189) = 71:40 ) h = 18:60 .

Digression.
With computer computation for aerial or space navigation, wherein which h (and
h0 ) might assume negative values, we can proceed in a somewhat di¤erent manner
than that employed in the LPB. Write

0
tan = tan cos l (30, repeated)

34
 tan 0 = tan 55 cos 312 = 0:9556. 0
= arctan (0:9556) ) 43:70 with the
0
positive sign (+) of uniquely determining that this 0 lies in the 1st trigonometric
quadrant (rather than in the 2nd ).
cos cos 55
From (31), we write cos 0 = 0
= = 0:793 4.
0
cos cos 43:70
= arcos(0:793 4) = 37:50 with the positive sign (+) of 0 uniquely determining
that 0 lies in the 1st trigonometric quadrant (rather than in the 2nd ).
0
= 110 43:70 = 66:30 .
0
cos ( ) = cos (66:30 ) = 0:4020.
From (33), cos = cos 0 cos ( 0
) = (0:7934) (0:4020) = 0:3190.
= arcos (0:3190) = 71:40 ) h = 18:60 .

Computations via the LPB Logarithmic Tables Round o¤ to a whole number.

1
A = 105 log sec 0
= 105 log = 10; 053
cos 37:50
1 1
B = 105 log sec (K d) = 105 log 0)
= 105 log = 39; 583
cos ( 0:40195
A + B = 10; 053 + 39; 583 = 49; 636

5 1 0:49636
(A + B) 10 = 0:49636 = log ) cos = 10 = 0:3189 > 0
cos
that is, 1st quadrant.

= arcos (0:3189) = 71:40

or, since sin h = sin (90 ) = cos , sin h = 0:3189 ) h = 18:60 (con…rma-
tion).

Azimuth calculation:
Again, from (25),
sin sin l sin 110 sin 312
sin Z = = = 0:7369
sin sin 71:40
Z = arcsin ( 0:7369) = 47:46 = 227:46 or 312:54

) that is, this evaluation indicates that Z lies in the 3rd or 4th trigonometric
quadrant.
The LPB predicts one of these values and provides directions for the navigator
to eliminate the quadrant ambiguity. However, in this example, we will provide
mathematical determination of the angle.

35
 From (16.3),
0:3190
z }| {
cos cos cos cos 110 cos 55 cos 71:4042
cos Z cos A = =
sin sin sin 55 sin 71:4042
= 0:6761 < 0

Z = arcos ( 0:6761) = 132:54 or 227:46

) that is, this evaluation indicates that Z lies in the 2nd or 3rd trigonometric
quadrant.
Zn = arctan 2 (cos Z; sin Z) = 227:46
End of LPB Numerical Example.

2.2 Dividing the Colatitude, Ageton’s Method

Figures 11A and 11B display the great circles

Figure 11A: Spherical Triangle, P East of M.

36
 Figure 11B: Spherical Triangle 2, P West of M.

This is the geometry used by Ageton in his "short-method" tables. In order

to avoid symbol congestion, primes (rather than, e.g., double primes) are attached
to some of the angle symbols. These must not be confused with the same
primed symbols employed in the section of this paper dealing with Ogura’s
method.
Again, for the upper spherical triangle in Figure 11A of Figure 11B, we make use
of Equations 25 and write...

sin sin Z = sin sin l

0 0
sin sin Z = sin sin l

Domains of the angles:

D( ) = [0; 180 ] ) ( ) : sin 0

D( ) = [0; 180 ] ) ( ) : sin 0

However, D(Z) = D(l) = [0; 360 ], so sin Z and sin l may be positive or negative.

But
sin Z sin sin Z 0 sin
= 0 and = 0
0
sin l sin sin l sin
) sign(sin Z) = sign(sin l):

37
 0 sin l
sin = sin
sin Z 0

0 sin l
sin = sin = sin jsin lj (36)
sin Z 0

cos = cos cos + sin sin cos A (16.3, repeated)

0 0 0 0
cos = cos cos + sin sin cos A0
0 0 0 0
cos = cos cos + sin sin cos (360 Z 0)
0 0 0 0
cos = cos cos + sin sin cos Z 0
0
Z 0 = 90 ) cos Z 0 = 0. Therefore, cos = cos cos 0
.

0 cos
cos = 0
(37)
cos

Derivation.

sin cos A = cos sin cos sin cos l (10.1, repeated)

0 0 0
sin cos A0 = cos sin cos sin cos l
But A0 = 90 or 270 ; therefore, cos A0 = 0;
/
0 0
cos sin = cos sin cos l
0
But from (37), cos = cos cos 0 , so
0 0 0
(cos cos 0 ) sin = cos sin cos l
0 0
cos sin = sin cos l (38)

cos = cos cos + sin sin cos l (23.3, repeated)

0
cos = (cos cos 0 ) cos + sin (cos 0
sin 0 )
0
cos = cos 0
(cos cos + sin sin 0 )
0 0
cos = cos cos ( ) (39)

End of calculations via Ageton’s Method.

38
Recapitulation The sequence of calculations used in the LPB (Sintiki Ogura’s
Method):
1. Use (29) sin 0 = sin jsin lj and 0
= arcsin (sin 0 ), assumed to lie in the
st
1 trigonometric quadrant.
cos
2. Use (31), cos 0 = 0
and 0 = arcos(cos 0 ) to uniquely calculate
cos
the angle 0 .
3. Use (33), cos = cos 0 cos ( 0
) and = arcos(cos ) to calculate
st
the angle , assumed to lie in the 1 trigonometric quadrant.

Another way of performing the calculations (This sequence not followed in

the LPB):
1. Use (30), tan 0 = tan cos l and 0 = arctan (tan 0 ) to uniquely calculate
the angle 0 . sign(cos 0 ) = sign(tan 0 ) .
cos
2. From (31), calculate cos 0 = 0
.
cos
3. Use (33), cos = cos 0 cos ( 0
) and = arcos(cos ) to uniquely
calculate the angle .

The sequence of calculations used in Ageton’s Method:

1. Use (36), sin 0 = sin jsin lj and 0
= arcsin (sin 0 ), assumed to lie in the
st
1 trigonometric quadrant.
cos
2. Use (37), cos 0 = and 0
= arcos(cos 0 ) to uniquely calculate
cos 0
the angle 0 .
0
3. Use (39), cos = cos 0 cos ( ) and = arcos(cos ) to calculate
st
the angle , assumed to lie in the 1 trigonometric quadrant.

39
2.3 Appendix
2.3.1 The Navigational Tetrahedron

Consider spherical triangles such as those of Figures 7A and 7B. Figure 12, below, is
representative of a tetrahedron in a general spherical triangle. We will use this …gure
to derive equations suitable for its description.

Figure 12: Details of the Navigational Tetrahedron.

The internal angles a; b; c (expressed in the lower case

above) subtend circular arcs. The surface angles A; B; C are angles between tangent
lines of intersecting circular arcs.
We derive equations on the basis of all angles less than 180 and greater than or
equal to zero (no unique directions).

m = OE n = OD; l = DE
p = AE; q = AD; R = OA
R = OB; R = OC

40
 The Spherical Trigonometry Law of Cosines

l2 = m2 + n2 2mn cos a
l2 = p2 + q 2 2pq cos A
m2 = R 2 + p2
n2 = R2 + q 2

Subtract the second equation above from the …rst equation:

0 = m2 p2 + n2 q2 2mn cos a + 2pq cos A

But m2 p2 = R2 and n2 q 2 = R2 , so

0 = R2 mn cos a + pq cos A

mn cos a = R2 + pq cos A
RR p q
cos a = + cos A
n m nm
In Figure 15, notice that the angles \OAD and \OAE are right angles.

R R q p
But cos b = ; cos c = ; sin b = ; sin c = :
n m n m
Then,
cos a = cos b cos c + sin b sin c cos A (40.1)
Likewise for the other two tangent lines, chords and angles. By similar reckon-
ing for the other vertices, lines and angles, we have the three "cosine equations" of
spherical trigonometry: a b; b c; c a.

cos b = cos c cos a + sin c sin a cos B (40.2)

cos c = cos a cos b + sin a sin b cos C (40.3)

0
Temporary symbols ; ; lying in the plane OED.

Problem: Determine whether CB ==DE. CB ==DE if and only if angle = .

In triangle OBC, the two sides OB and OC are both equal to R ; Hence, the
same angle on both sides of the triangle. In triangle ODE, we may not assume that

41
 R R
m = n; m = and n = . But angle b is not necessarily equal to angle c, so
cos c cos b
m is not necessarily equal to n and thus, angle 0 is not necessarily equal to angle .
Consequently, CB is not necessarily parallel to DE, i.e., l.

The Spherical Trigonometry Law of Sines

From (40.1),
cos a cos b cos c
cos A =
sin b sin c
(cos a cos b cos c)2
sin2 A = 1 cos2 A = 1
sin2 b sin2 c

2 sin2 b sin2 c (cos a cos b cos c)2

sin A =
sin2 b sin2 c
(1 cos2 b) (1 cos2 c) (cos a cos b cos c)2
=
sin2 b sin2 c

2 sin2 b sin2 c (cos a cos b cos c)2

sin A =
sin2 b sin2 c
1 cos b cos2 c + cos2 b cos2 c (cos2 a 2 cos a cos b cos c + cos2 b cos2 c)
2
=
sin2 b sin2 c
1 cos2 a cos2 b cos2 c + 2 cos a cos b cos c
sin2 A =
sin2 b sin2 c
sin2 A 1 cos2 a cos2 b cos2 c + 2 cos a cos b cos c
=
sin2 a sin2 a sin2 b sin2 c
p
sin A 1 cos2 a cos2 b cos2 c + 2 cos a cos b cos c
=
sin a sin a sin b sin c
sin B sin C
When we do likewise for and , the right hand side of each of these
sin b sin c
equations is equal to
p
1 cos2 a cos2 b cos2 c + 2 cos a cos b cos c
.
sin a sin b sin c

From this we have what are known as the Law of Sines for Spherical Trigonometry:

sin A sin B sin C

= = (41)
sin a sin b sin c
Equations 40.1, 40.2 and 40.3 express cos a, cos b and cos c as trigonometric func-
tions of angles a, b , c, A, B , C. Similarly, we would like to create some kind of

42
converse relationship to these equations, expressing cos A, cos B and cos C as trigono-
metric functions of angles a, b , c, A, B , C. We refer to this "converse relationship"
as the second spherical law of cosines. Three instances of this law are presented below,

cos A = cos B cos C + sin B sin C cos a (42.1)

cos B = cos C cos A + sin C sin A cos b (42.2)

cos C = cos A cos B + sin A sin B cos c (42.3)

We have previously discussed this law at length in the main text:
Suppose that we introduce the following replacements of angle symbols:
(a; A) ( ; l) for the circular arc of length R .
(b; B) ( ; Z) for the circular arc of length R .
(c; C) ( ; ) for the circular arc of length R .
sin A sin B sin C sin l sin sin Z
Then, = = is replaced by = = or
sin a sin b sin c sin sin sin
sin LHA sin B sin A
= = .
sin sin sin
Hence, we see that sin sin A = sin sin l, which we had earlier derived in the
original rotation matrix derivation as equation 10.2.
Similarly, equations 40.1, 40.2, 40.3 are replaced by equations 10.3, 16.3 and 26.3
respectively (repeated below).
cos = cos cos + sin sin cos l
cos = cos cos + sin sin cos A
cos = cos cos + sin sin cos

Many authors derive the second spherical law of cosines by considering "polar
triangles", which are discussed in some of the texts appearing in the bibliography. We
will not address those concepts here, except to mention that the equations expressing
the second law can be obtained by making the following replacements in the spherical
law of cosines:
A 180 a a 180 A
B 180 b and b 180 B
C 180 c c 180 C
However, these replacements alone do not constitute mathematical proofs.
All of the derivations in this appendix are based upon the assumption that the
domains of all surface angles, D(A) = D(B) = D(C) = [0; 180 _]. However, based
upon arguments in the text derived via rotation matrices, we see that the domain of
the surface angles is [0; 360 ), provided that we adhere to the interpretation of those
angles as given in the text.

43
2.4 Bibliography
1. Cotter, Charles H., The Complete Nautical Astronomer; Hollis & Carter;
London, 1969. ISBN 0-370-00230-X.

2. Cotter, Charles H., "A Study of a Species of Short-Method Table",

http://www.cambridge.org/core.

3. Hann, James, The Elements of Spherical Trigonometry; a new edition, revised

and corrected by Charles H. Dowling; Virtue Brothers & co., London, 1866.

4. Murray, Daniel A., Spherical Trigonometry for Colleges and Secondary

Schools; Longmans, Green, and Company, London, 1908.

5. Newcomer, K.E., "Navigation Applications of the HP-65 Calculator",

Navigation, Journal of the Institute of Navigation, Vol. 22, No.2, Summer
1975.

6. Palmer, Claude Irwin, and Leigh, Charles Wilber, Plane and Spherical
Trigonometry, fourth edition; McGraw-Hill Book Company, Inc.; New York
and London, 1934.

7. Schlereth, Hewitt, Commonsense Celestial Navigation, Henry Regnery

Company, Chicago, 1975; ISBN 0-8092-7219-9 (paper).

8. Todhunter, Isaac, Spherical Trigonometry, For the Use of Colleges and Schools,
5th edition; Macmillan and Company, London, 1886.

9. Umland, Henning, A Short Guide to Celestial Navigation,

2006. www.celnav.de.

10. Weems, P.V.H. and Lee, C.V., Marine Navigation, 2nd edition, Weems System
of Navigation, Inc., Annapolis, Maryland, 1958. Library of Congress Card
No. 58 - 10342.

11. Weems, P.V.H, Line of Position Book, fourth edition, Weems System
of Navigation; Annapolis, 1943.

12. Wentworth, George and Smith, David Eugene, Plane and Spherical
Trigonometry; Ginn and Company (The Athenaum Press, U.S.A.), 1915.

13. http://www.siranah.de/ and http://www.siranah.de/html/sail040u.htm.

[Ageton’s method].

End

44