Arabic Sciences and Philosophy, vol. 5 (1995) pp.

145-193
Copyright © 1995 Cambridge University Press

IBN AL-HAYTHAM'S UNIVERSAL SOLUTION

FOR FINDING THE DIRECTION OF THE QIBLA
BY CALCULATION
AHMAD S. DALLAL

INTRODUCTION

This paper includes an edition, translation of, and commentary

on a treatise by Abu 'All al-Hasan b. al-Hasan b. al-Haytham on
finding the azimuth of the qibla by calculation. The treatise has
never been studied before in modern times.
Ibn al-Haytham was born in 965 in Basra. He spent most of
his professional life in Baghdad and later in Cairo, where he
died in 1040. He wrote on many subjects, but the extant works
for which he was mostly known are on physics, astronomy,
mathematics, and optics.1
The qibla is the sacred direction which Muslims face during
prayer and other ritual acts.2 Finding the direction of the qibla
was a favorite problem for medieval astronomers, and it was
treated in separate works, in sections of zijes and astronomical
handbooks, or in manuals for time-keeping. Some of the great-
est minds among the Muslim astronomers turned their atten-
tion to this problem and devised for this purpose solutions of
1
On the life and work of Ibn al-Haytham see the latest analysis in R. Rashed, Les
mathematiques infinitesimales du IXe au XIe siecle. Vol II: Ibn al-Haytham (London,
1993). In the introduction to this work, Rashed resolves a long standing ambiguity in
the biographical studies on Ibn al-Haytham, and proves that there are in fact two his-
torical figures by this name; these are al-Hasan ibn al-Hasan, and Muhammad ibn al-
Hasan. The former is the most famous of the two, and is the author of the work stud-
ied here, while the later is a philosopher and physician. For an earlier account of Ibn
al-Haytham's life and work see A. I. Sabra, "Ibn al-Haytham," Dictionary of Scientific
Biography, 18 vols. (New York, 1970-90), vol. 6 (1972), pp. 189-210. For lists of his
works and available manuscripts, see F. Sezgin, Geschichte des arabischen
Schrifttums, 9 vols. (Leiden, 1967 Onwards), especially Band V, Mathematik (1974),
pp. 358-73, and Band VI (1978), Astronomie, pp. 251-61.
2
On the sacred direction see D. A. King, "Kibla," Encyclopaedia of Islam, second
edition, 6 vols. to date (Leiden, 1960 to present) vol. V, pp. 82-8. See also D. A. King,
"The sacred direction in Islam: A study of the interaction of religion and science in
the Middle Ages," Interdisciplinary Science Reviews, 10 (1985): 315-28.
146 AHMAD S. DALLAL
considerable sophistication. The development of methods
reflected a development in the use of mathematical techniques;3
the earliest and least complex methods were approximations;
next, solid geometry was used to accurately calculate the direc-
tion of the qibla* various analemmas were then employed to
solve the problem of the qibla graphically, by orthogonal projec-
tion of the space figure onto a single plane, thence rotating
oblique planes 5down into it to obtain the true dimensions of
their contents; finally, using spherical trigonometry, specifi-
cally the theorem of Menelaos or the simpler sine theorem, the
direction of the qibla could be derived by using entities on the
surface of the sphere.6 Moreover, numerous tables giving the
3
On the standard approximation method of al-Battani see King, "Kibla," p. 84; also
see D. King, "Al-Khalili's qibla table," Journal of Near Eastern Studies, 34.2 (1975):
81-121, p. 82. For discussion of three approximate computational methods and one
approximate cartographic construction see D. King, "The earliest Islamic mathemat-
ical methods and tables for finding the direction of Mecca," Zeitschrift fur Geschichte
der Arabisch-Islamischen Wissenschaften, 1 (1986): 97-109. For tables based on
approximate methods see King, "Al-Khalili's qibla table," pp. 120-2, and King,
"Earliest," pp. 108-12.
4
For exact solutions of the problem of the qibla using rectilinear configurations
inside the sphere see J. Ali, The Determination of the Coordinates of Cities: al-
Biruni's Tahdid al-Amakin (Beirut, 1967), pp. 243-52; E. S. Kennedy, A
Commentary upon al-Biruni's Kitab Tahdid al-Amakin (Beirut, 1973), pp. 202-9; and
King, "Earliest," pp. 112-15.
5
On different analemmas used for solving the qibla problem see E. S. Kennedy and
Y. Id, "A letter of al-Blrum: Habash al-Hasib's analemma for the qibla," in E. S.
Kennedy, Colleagues and Former Students, Studies in the Islamic Exact Sciences
(Beirut, 1983), pp. 621-9; Ah', Biruni's Tahdid, pp. 255-6; Kennedy, Commentary upon
al-Biruni's Tahdid, pp. 209-11; For a comparison of four analemmas including, in addi-
tion to the above two by Habash and Biruni, one analemma from Biruni's Al-Qanun al-
Mas'udi and one by Ibn al-Haytham see J. L. Berggren, "A comparison of four
analemmas for determining the azimuth of the qibla," Journal of the History of Arabic
Science, 4 (1980): 69-80. Also, on the analemmas of Habash and Ibn al-Haytham, see
King, "Kibla," p. 85, and King, "Earliest," pp. 115-18. For tables based on analemmas
see King, "Kibla," p. 87, and King, "Al-Khalili's qibla table," pp. 101-8,110-11.
6
For spherical trigonometric methods see Biruni's first and fifth methods in Ali,
Biruni's Tahdid, pp. 241-3, 252-5, and Kennedy, Commentary upon al-Biruni's
Tahdid, pp. 198-200, 211-14. For a comparison and discussion of the development of
such methods in the works of Habash al-Hasib (9th century), Ibn Yunus, Abu al-
Wafa' al-Buzjani, al-Quhi, Kushyar ibn Labban, al-Biruni, and the anonymous author
of the Zij al-Shamil (10th and 11th centuries), and Jamshid al-Kashi (15th century),
see J. L. Berggren, "On al-Biruni's 'Method of the zijes' for the qibla," Proceedings of
the 16th International Congress for the History of Science (Bucharest, 1981), pp. 237-
45, and J. L. Berggren, "The origins of al-Biruni's 'Method of the Zijes' in the theory
of sundials," Centaurus, 28 (1985): 1-16. Also on the methods of al-Biruni and al-
Nayrizi see King, "Kibla," pp. 85-6. For the earnest exact method using spherical
trigonometry see King, "Earliest," pp. 115-18. For tables based on such methods see
King, "Kibla," pp. 87-8; King, "Earliest," pp. 118-29; and King, "Al-Khalili's qibla
table."
 IBN AL-HAYTHAM'S UNIVERSAL SOLUTION 147
direction of the qibla for different localities were computed on
the bases of each of the above methods.7
Ibn al-Haytham composed three different treatises on finding
the direction of the qibla by three different methods:
1) "Qawl fi samt al-qibla bi-al-hisab" (On (Finding) the
Azimuth of the Qibla by Calculation).8 This is the work that
forms the subject of the present study.
2) "Qawl (or maqala) fi istikhraj samt al-qibla" (On the
Determination of the Azimuth of the Qibla).9
3) A third treatise listed in Ibn al-Haytham's autobiography is
entitled "Maqala fi istikhraj samt al-qibla fijamV al-maskuna
bi-jadawil wada'tuha wa lam urid al-burhdn 'aid dhalik" (On
Finding the Azimuth of the Qibla in all the Inhabited (Lands)
by (Using) Tables Which I Compiled Without Giving the Proof).
This treatise is apparently lost.10
Several studies of the mathematics of the11
qibla determination
were produced in the past three decades. With the exception of
the references by King,12 however, the two extant treatises by
Ibn al-Haytham were often confused, and as a result only his
analemma method was given proper consideration and treat-
ment. In Band V (Mathematik), F. Sezgin lists four copies of the
first treatise, "Qawlfisamt al-qibla bi-al-hisdb," in the libraries
of Istanbul and Tehran. For this study I have not been able to
consult any of these four copies, but have used two other manu-
scripts; the first is Berlin Oct. 2970/1 (fols. 4 r -ll v ), copied by
Qadlzadeh. This copy is mistakenly listed in Sezgin's Band V
(Mathematik) under the second treatise "Qawl fi istikhraj samt
al-qibla," but the error is corrected in Band VI (Astronomie).
7
See, in addition to references listed in the above footnotes, King, "Earliest," pp.
130-41. It should be noted that the working inside the sphere, analemmas, and work
on the surface of the sphere all started in Hellenistic times, but these methods were
greatly augmented in the Islamic Middle Ages.
8
For a list of available manuscripts see Sezgin, Geschichte, Band V, # 15, p. 368,
and Band VI, # 18, p. 259. Also for a reference to this treatise see Sabra, "Ibn al-
Haytham," p. 205.
9
For a list of available manuscripts see Sezgin, Geschichte, Band V, # 16, p. 368,
and Band VI, # 19, p. 259.
10
For an edition of Ibn al-Haytham's autobiography see Anton Heinen, "Ibn Al-
Haytams Autobiographie In Einer Handschrift Aus Dem Jahr 556 H./1161 A.D.," Die
islamische Welt zwischen Mittelalter und Neuzeit: Festschrift fur Hans Robert Roemer
zum 65. Geburtstag (Beirut, 1979), pp. 254-77. For a reference to the third treatise
see p. 263 of the autobiography. For the second treatise see p. 276.
11
See footnotes 3-7 above.
12
See King, "The sacred direction," p. 317, and King, "Earliest," p. 116.
148 AHMAD S. DALLAL
Therefore, the manuscript Berlin, Oct. 2970/1 is not a copy of
Ibn al-Haytham's "istikhraj samt al-qibla," which is a solution
of the problem of the qibla by using an analemma construction,
and which was studied by Schoy in the first13 serious modern
examination of the mathematics of the qibla., Furthermore, in
his biography of Ibn al-Haytham, Sabra lists only one work on
the qibla, namely the "Qawl fi samt al-qibla bi-al-hisdb," and
mistakenly indicates that this is the same work translated into
German by Schoy.14 The confusion may have resulted from the
opening statement in the second treatise "Qawl fi istikhraj
samt al-qibla," which reads:
Kunna allafna maqala fi istikhraj samt al-qibla fijami' al-mawadi' min al-
ard shamaliha wajanubiha bi-tariq al-hisab wa al-barahin...
We have already composed a treatise on the extraction of the azimuth of
the qibla in all northerly and southerly localities on the earth using compu-
tational methods and proofs...

Ibn al-Haytham then goes on to state that he will next compose

a short treatise for the northerly localities only, and will use a
non-computational method.15 The cataloguer of this manuscript
may have noticed the clause "istikhraj samt al-qibla ... bi-tariq
al-hisab" in the above statement and assumed, without further
examination, that the two treatises are the same.
The second manuscript consulted for this work is not listed in
Sezgin. It is MS Cairo, Dar al-Kutub K 3823/3, (fols. 14V-18V),
copied ca. 900 AH (ca. 1500 AD) from a copy by the hand of
Qadizadeh.16
I know only of one reference to this work by Ibn al-Haytham
in the later scientific literature, namely in the treatise on
instruments by the fourteenth-century Aleppo astronomer Ibn
al-Sarraj, who says:

13
See C. Schoy, "Abhandlung des al-Hasan ibn al-Hasan ibn al-Haitam (Alhazen)
uber die Bestimmung der Richtung der Qibla," Zeitschrift der Deutschen
Morgenlandischen Gesellschaft, 75 (1921), pp. 242-53. Also see C. Schoy, "Kibla,"
Encyclopaedia of Islam, first edition, 4 vols. (Leiden, 1913-34) vol. II, pp. 988-9.
14
See Sabra, "Ibn al-Haytham," p. 205.
15
See Schoy, "Abhandlung," p. 244.
16
For a catalogue reference to this manuscript see D. A. King, A Catalogue of The
Scientific Manuscripts in the Egyptian National Library (in Arabic), (Cairo, 1981), p.
317. Also see D. A. King, A Survey of the Scientific Manuscripts in the Egyptian
National Library (Malibu, Ca., 1985), # B.77-(3.3.1). On Qadizadeh see Hamit
Dilgan, Dictionary of Scientific Biography, vol. 11 (1975), pp. 227-9.
 IBN AL-HAYTHAM'S UNIVERSAL SOLUTION 149
Abu al-Hasan ibn al-Hasan ibn al-Haytham has done a brilliant job in solv-
ing this (qibla) problem by articulating it in sixteen figures (accompanied) by
proofs.17

Ibn al-Sarraj's reference to the ingenuity of the solution of Ibn

al-Haytham is accurate. So far, however, the importance of this
solution has not been recognized in the literature on either Ibn
al-Haytham or the mathematics of the qibla. In his comparative
study of the exact spherical-trigonometric methods used for
solving the problem of the qibla, Berggren18 examines, among
others, the methods of al-Biruni19 (11th century), al-Quhi20 (10th
century), Abu al-Wafa' al-Buzjani21 (10th century), Kushyar ibn
Labban22 (10th century), the anonymous author of the Zij al-
Shamil,23 and Jamshid al-Kashi24 (15th century). Berggren uses
three criteria for comparing these methods, namely, the values of
the computed auxiliary arcs, the extent to which the possibilities
of the positions of Mecca relative to various localities are elabor-
ated, and the names given to the various computed auxiliary arcs.
The methods developed in the above works involve computing
a series of auxiliary arcs culminating in the arc which measures
the azimuth of the qibla. Although the methods of computation
may differ25, the arcs calculated in these methods are closely

17
Ibn al-Sarraj is quoted in King, "Sacred direction," p. 317. The text, taken from
MS Dublin Chester Beatty 4833, reads: "Laqad abda'a Abu al-Hasan ibn al-Hasan
ibn al-Haytham fi tahrir hadhihi al-mas'ala idh sawwaraha fi yet' waw (16) shaklan
mubarhanan." On Ibn al-Sarraj see King, Survey, # C26.
18
See Berggren, "Biruni's 'Method'," pp. 237-45, and Berggren, "Origins," pp. 1-16.
19
On al-Biruni see E. S. Kennedy, "Al-Biruni," Dictionary of Scientific Biography,
vol. 2 (1970), pp. 147-58.
20
On al-Quhi see Y. Dold-Samplonius, "Al-Quhi," Dictionary of Scientific
Biography, vol. 11 (1975), pp. 239-41.
21
On Abu al-Wafa' see A. P. Youschkevitch, "Abu al-Wafa' al-Buzjani," Dictionary
of Scientific Biography, vol. 1 (1970), pp. 39-43.
22
On Kushayr see A. S. Saidan, "Kushyar ibn Labban," Dictionary of Scientific
Biography, vol. 7 (1973), pp. 531-3.
23
On this and other zijes see E. S. Kennedy, "A survey of Islamic astronomical
tables," Transactions of the American Philosophical Society, 46 (1956): 123-77.
24
On al-Kashi see A. P. Youschkevitch and B. A. Rosenfeld, "Al-Kashi," Dictionary
of Scientific Biography, vol. 7 (1973), pp. 255-62.
25
For example, Biruni proposes in his al-Qanun al-Mas'udi a refined computa-
tional method in which he uses the rule of four quantities, and the law of sines for
spherical triangles instead of the cumbersome application of the theorem of
Menelaos. See King, "Kibla," p. 86, and Berggren, "Origins," pp. 14-15. On the rule
of four and other developments in Islamic trigonometry see Kennedy, "The history of
trigonometry," in Kennedy, Colleagues and Students, Studies in the Islamic Exact
Sciences, pp. 3-29.
150 AHMAD S. DALLAL
related, and the arcs calculated by Ibn al-Haytham are no
exception.
The second criterion used by Berggren is to examine the
"elaboration of the method to deal with a wider variety of cases
of the position of Mecca relative to the locality in question."26
Wlaereas tYve earliest exact metYio&s Yiad no sucYi refinements,
the question of the direction in which the azimuth of Mecca
should be measured is already considered in the Zij al-Shamil,
and in works by BirunI and Kushayr ibn Labban.27
Furthermore, Berggren argues that "this elaboration reached
its culmination in the works of Jamshid al-Kashi... who, in his
Zij al-Khaqanl, elaborated and supplemented the method to
deal with worshippers at any location on the earth's surface and
thereby created from the "method of the zijes" a universal solu-
tion of the qibla problem."28 The present work by Ibn al-
Haytham, however, provides a "universal" solution of the prob-
lem of the qibla some four centuries prior to the work of al-
Kashi.
Berggren also points out that whereas some of the examined
works use the names "modified longitude/latitude" (tul
mu'addal, 'ard mu'addal), or the "correction of longitude/
latitude" (ta'dll al-tul, ta'dll al-'ard), others use 29
the names "the
first quantity" and "the second quantity," etc. According to
Berggren, the mathematical procedures, as well as the language
used in the first set of solutions, indicate that these solutions of
the qibla problem employed a double transformation of coordi-
nates, and the measurement of the longitude and the latitude
with respect to the meridian. Berggren thus concludes that the
roots of the method of the zijes is in the theory of sundials.30
King has noted, however, that the language and computations
of the earliest exact method are independent of the methods
used in traditional sundial theories.31 The present work by Ibn
al-Haytham, seems to combine elements of the above two tradi-
tions, such that the computed arcs are measured with respect to
the meridian, but the names "the first arc," and "the second
26
See Berggren, "Origins," p. 16.
27
See Berggren, "Birunl's 'Method'," p. 245, and Berggren, "Origins," pp. 5-6, 8-9.
28
See Berggren, "Origins," p. 16.
29
For the earliest extant method where the arcs are called "the first quantity," etc.,
see King, "Earliest," pp. 112-15.
30
See Berggren, "Origins," pp. 11-14.
31
See King, "Earliest," p. 117.
 IBN AL-HAYTHAM'S UNIVERSAL SOLUTION 151
arc," etc., rather than "modified" or "corrected" longitudes and
latitudes are used.
Ibn al-Haytham's computational method for finding the direc-
tion of the qibla is thus important for two reasons: it represents
an important juncture in the history of the development of the
mathematics of the qibla, and it augments our knowledge of the
contributions of one of the most important scientists of
medieval Islam. It is surprising that in the computational
method of this treatise Ibn al-Haytham confines himself to the
use of the old-fashioned Menelaos theorem, and that he disre-
gards the developments in trigonometry that took place
between the ninth and the eleventh centuries. Ibn al-Haytham,
however, was not unique in this regard. His contemporary al-
Biruni, for example,32
solved the qibla problem by using several
different methods, including one which employs the more ele-
gant sine theorem. Biruni also solved the qibla problem using
the Menelaos theorem. It is apparent, therefore, that the persis-
tence on using the Menelaos theorem does not reflect the
author's incompetence in the more developed techniques of
trigonometry. My tentative explanation of Ibn al-Haytham's
use of this cumbersome theorem in the earliest extant compre-
hensive mathematical treatment of the qibla problem is that
this work addressed a broad audience which had some knowl-
edge of astronomy and trigonometry, but was not necessarily
versed in the latest technical developments in these fields.
Should this hypothesis be true, the present work by Ibn al-
Haytham may shed some light on the range of scientific activity
and its dissemination in medieval Islamic societies.
Critical apparatus:
In the margins of the edited Arabic text I have marked the MS
sigla and folio numbers of the two manuscripts used. I have
employed square brackets to indicate the sequential paragraph
numbers which are obviously not part of the original text. As
for the English translation and commentary, the paragraphs
of each chapter were numbered following the numbering of
the Arabic text. These numbers appear inside square brackets.
Words in parentheses have been added to the translation in
order to conform to English idiom as much as possible, and to
clarify the meaning. Thefigureswere reconstructed and inserted
32
On Biruni's methods see footnotes 4-6 above.
152 AHMAD S. DALLAL
in the Arabic original and the translation. Figures that do not
appear in the original have been inserted in the commentary.

TRANSLATION

[1] In the name of God, the Merciful, the Compassionate.

[2] An essay on (finding) the azimuth of the qibla by calcula-
tion, (written) by al-Hasan ibn al-Haytham.
[3] The azimuth of the qibla is that direction which, when faced
by a person, it will be as if he is looking towards the diameter of
the world passing through the ka'ba. The (person's) line of
vision along this azimuth will fall in the plane of the great circle
which passes through his local zenith and through the zenith of
the ka'ba. This azimuth, I mean, the azimuth of the qibla, is
demarcated by the straight line which is the intersection
between the horizon of the locality at which the azimuth is
required, and the (plane of the) great circle which passes
through the pole of that horizon and through the pole of the
horizon of the ka'ba.
[4] This line is determined by finding the longitudes and lati-
tudes of the two localities; I mean by the two localities: that of
the ka'ba, and the locality at which we want to extract the
azimuth. So if we wanted to determine the azimuth for any
desired locality we find the longitude and latitude of the locality,
and the longitude and latitude of Mecca. We then consider (the
following cases): if the two longitudes were equal, I mean, if
both localities fell on the same meridian circles, then the
azimuth line will be the meridian line. If the locality was north
of the locality of the ka'ba, I mean, if the northerly latitude of
the locality was more than the latitude of Mecca, then the
azimuth will be on the southern side of the meridian line. If,
however, the locality was south of the locality of the ka'ba, then
the azimuth will be on the northern side of the meridian line. If
the difference between their longitudes was a semicircle, and if
(the locality) was diametrically opposite the position of the
ka'ba, I mean, if it (the locality) was south of the equator on the
(great) circle which passes through the zenith of the ka'ba, and
if its southerly latitude was equal to the latitude of Mecca -
which is (true) if the difference between their latitudes was also
a semicircle - then each of the two sides of the meridian line will
be the azimuth of the qibla for that locality. If, however, it (the
 IBN AL-HAYTHAM'S UNIVERSAL SOLUTION 153
locality) was not diametrically opposite (the ka'ba), and if the
shorter of the two arcs (of the great circle) that passes through
both of them was the one which passes through the north pole,
I mean, if the locality was north, on, or south of the equator,
such that its latitude was less than the latitude of Mecca, then
the azimuth of the qibla will be on the northern side of the
meridian line. If the shorter (arc) was the one which passes
through the south pole, I mean, if the locality was south of the
equator, and if the latitude (of the locality) was more than the
latitude of Mecca, then the azimuth will be on the southern side
of the meridian line. If, (on the other hand,) the two longitudes
were different, and if the difference between them was less than
a semicircle, then the azimuth is determined by extracting the
arc of the horizon (circle which falls) between the meridian line
and the (great) circle that passes through the zeniths of the two
localities. This arc is called the azimuth arc.
[5] The determination of this arc is as I describe (in what fol-
lows): if the difference between the two longitudes was less or
more than a quadrant, but less than a semicircle, then we mul-
tiply the sine of the difference between the two longitudes by
the sine of the complement of the latitude of Mecca. We then
divide the product by sixty, if the sine table was calculated for a
diameter of one hundred and twenty (units). If (the sine table
was) calculated for some other number, then we divide by half
of this number. We enter the result (of the above computation)
into the sine table and find its arcsine, which we call the first
arc. We then subtract it from ninety, compute the sine of the
remainder, and save it. Next we, multiply the sine of the lati-
tude of Mecca by sixty, and divide it by the sine we had saved;
we then find the arcsine of the result, and call the outcome the
second arc.
[6] Next we consider (the following cases): If this (second) arc
was equal to the latitude of the locality at which the azimuth is
required, and if the latitude was north of the equator, and the
difference between the two longitudes was less than a quadrant,
then the azimuth of the qibla will fall on the east-west line - if
the locality was east of the locality of the ka'ba, then (the
azimuth will be) on the west side (of the east-west line), and if
the locality was to the west, then (the azimuth will be) on the
east side. If, however, this (second) arc was equal to the latitude
(of the locality), and if the latitude was south of the equator,
and the difference between the two longitudes was more than a
154 AHMAD S. DALLAL
quadrant, then the azimuth of the qibla will be on the east-west
line, also to the same side on which the ka'ba falls.
[7] If this arc, which we designated the second was equal to the
complement of the latitude (of the locality), and if the latitude
was northerly, and the difference between the two longitudes
was greater than a quadrant, then the arc which we designated
the first will be the azimuth arc, and it will be northerly. If the
second arc was equal to the complement of the latitude, and if
the latitude was southerly, and the difference between the two
longitudes was less than a quadrant, then the first arc will be
the azimuth arc, and it will be northerly.
[8] If the latitude was northerly, and the difference between
the two longitudes was greater than a quadrant, we measure
the second arc against the complement of the latitude: if it was
smaller than the complement of the latitude, then we subtract
it from this complement, find the sine of the remainder, and call
it the auxiliary sine {al-jayb al-muhassal, literally the resultant
sine). If it was greater than the complement of the latitude,
then we subtract the complement from it, find the sine of the
remainder, and this will be the auxiliary sine. If the second arc
was greater or smaller than the latitude, and if the latitude was
northerly, and the difference between the two longitudes was
less than a quadrant, then we add the second arc to the comple-
ment of the latitude. We then find the sine of the sum, and this
will be the auxiliary sine.
[9] If the latitude was southerly, and the difference between
the two longitudes was greater than a quadrant, and if the sec-
ond arc was greater or smaller than the latitude, then we add
the second arc to the complement of the latitude. We then find
the sine of the sum, and this will be the auxiliary sine. If the lat-
itude was southerly, and the difference between the two longi-
tudes was less than a quadrant, then we measure the second arc
against the complement of the latitude: if it was less than the
complement of the latitude, then we subtract it from this com-
plement, find the sine of the remainder, and this will be the aux-
iliary sine. If, however, it was greater than the complement of
the latitude, then we subtract the complement from it, find the
sine of the remainder, and this will be the auxiliary sine.
[10] If the locality was on the equator, and the difference
between the two longitudes was less or more than a quadrant,
then we subtract the second arc from ninety (degrees), find the
sine of the remainder, and this will be the auxiliary sine.
 IBN AL-HAYTHAM'S UNIVERSAL SOLUTION 155
[11] Next, we multiply the auxiliary sine by the sine which we
had saved, and divide the product by sixty; we then enter the
result into the sine table and find its arcsine; we subtract this
arc from ninety, find the sine of the remainder, and save it.
Next we multiply the sine of the first arc by sixty, and divide it
by the last sine which we had saved. We enter the product into
the sine table and find its arcsine. The result will be the
azimuth arc.
[12] If, (on the other hand,) the difference between the two lon-
gitudes was a quadrant, and the latitude was northerly or
southerly, then we multiply the sine of the latitude of the local-
ity at which the azimuth is required by the sine of the latitude
of Mecca, and divide the product by sixty; we then find the arc-
sine of the result and subtract it from ninety, find the sine of
the remainder, and save it. Next, we multiply the sine of the
complement of the latitude of Mecca by sixty, and divide it by
the sine which we have just saved. We find the arcsine of the
result, and this will be the azimuth arc.
[13] If the difference between the two longitudes was a quad-
rant, and the locality was on the equator, then the azimuth arc
will be the complement of the latitude of Mecca.
[14] Once we determine this (azimuth) arc we consider (the fol-
lowing cases): if the latitude of the locality at which the azimuth
is required was northerly, and the difference between the two
longitudes was not less than a quadrant, then the azimuth arc
will be to the north of the locality. If, (on the other hand,) the
latitude was northerly, and the difference between the two lon-
gitudes less than a quadrant, then we consider the arc which we
designated the second: if it was greater than the latitude of the
locality, then the azimuth arc will be northerly. If, however, the
second arc was smaller than the latitude of the locality, then the
azimuth arc will be southerly. If the latitude was southerly, and
the difference between the two longitudes was not greater than
a quadrant, then the azimuth arc will be northerly. If, (on the
other hand,) the latitude was southerly, and the difference
between the two longitudes was greater than a quadrant, then
we consider the second arc: if it was smaller than the latitude,
then the azimuth arc will be southerly. If, however, it was
greater than the latitude, then the azimuth arc will be
northerly. Next we consider (the following cases): if the locality
at which the azimuth is required was east of the locality of the
ka'ba, then the azimuth arc will be westerly. This will be (true)
156 AHMAD S. DALLAL
if the longitude of the locality was more than the longitude of
Mecca, and the longitude was measured from the west side, or if
the longitude of Mecca was more than the longitude of the local-
ity, and the longitude was measured from the east side. If, (on
the other hand,) the locality was west of the locality of the
ka'ba, then the azimuth arc will be easterly. This will be (true)
if the longitude of the locality was more than the longitude of
Mecca, and the longitude was measured from the east, or if the
longitude of Mecca was more than the longitude of the locality,
and the longitude was measured from the west side.
[15] Once we are done with all of these (considerations), we
adjust a plane parallel to the horizon, draw a circle in it, and
draw the meridian line which passes through its center. We
divide the circle into quadrants. Then, having determined
whether the direction of the azimuth arc is northerly or
southerly, and whether it is easterly or westerly, we consider
the quadrant which (falls) in this direction. We divide this quad-
rant into ninety equal degrees. Of these (ninety degrees) we sep-
arate a portion measured from the meridian line, and equal to
the value which we extracted by the above mentioned calcula-
tions and called the "azimuth arc." We then join by a straight
line the (point) marking the end of the degrees (of the azimuth
arc) and the center of the circle. This will be the required
azimuth. (Alternatively,) we can, if we wish, superpose over the
plate which is parallel to the horizon (another) plate in which a
quadrant is divided into ninety degrees. Let the diameter (of the
superposed plate) coincide with the meridian line. Next, place
the calibrated quadrant on the side where the azimuth arc falls,
and (measure), from the extremity of the diameter, an amount
of degrees equal to the azimuth arc, which was extracted by cal-
culation. We join by a straight line the resulting point and the
center of the plate. We then issue, in the plane parallel to the
horizon, a straight line which extends along the line (con-
structed above). This will be the azimuth line.
[16] Now, let us present the proof for the verity of these calcu-
lations. Let the meridian circle for the locality at which the
azimuth is required be ABG [Fig. A]; the zenith: point B; the
horizon: circle AEG; the equator circle: ZDE, with point L as its
north pole; and the meridian circle of Mecca: LHM. Let the
zenith of the ka'ba be H. Also, let the difference between the
two longitudes be less or more than a quadrant, and let the
azimuth (samtiyya) circle which passes through the zeniths of
 IBN AL-HAYTHAM'S UNIVERSAL SOLUTION 157

Figure A

the two localities be circle BHT. Arc AT will then be the

azimuth arc. Imagine the great circle passing through the
points E (and) H, which is circle EHK. Since the two arcs EK
(and) LM intersect between the two arcs LD (and) DE at point
H, the ratio of the sine of arc DE to the sine of arc DM is com-
pounded of the ratio of the sine of arc EK to the sine of arc KH,
and the ratio of the sine of arc HL to the sine of the arc LM. But
each of ED (and) EK equals a quadrant, since point E is the pole
of circle ABG; therefore, the ratio of the sine of arc ED to the
sine of arc DM is compounded of the ratio of the sine of arc ED
to the sine of arc KH, and the ratio of the sine of arc HL to the
sine of arc LM. Also, the ratio of the sine of arc ED to the sine
of arc DM is compounded of the ratio of the sine of arc ED to the
sine of arc KH, and the ratio of the sine of arc KH to the sine of
arc DM. So the ratio of the sine of arc KH to the sine of arc DM
is as the ratio of the sine of arc HL to the sine of arc LM. So, if
we multiply the sine of arc DM by the sine of arc HL, and divide
by the sine of arc LM, we will obtain the sine of arc KH. But the
sine of arc DM is the sine of the difference between the two lon-
gitudes, arc HL is the complement of the latitude of Mecca, and
arc LM is a quadrant whose sine is sixty. So, if we multiply the
sine of the difference between the two longitudes by the sine of
158 AHMAD S. DALLAL
the complement of the latitude of Mecca, and divide (the prod-
uct) by sixty, we will obtain the sine of the arc HK. Now if we
find the arcsine (of this last quantity), then the (resulting) arc
will be arc KH, which we designated the first arc. If then we
subtract (this arc) from ninety, then the remainder will be arc
EH. Thus, arc EH is the one whose sine we have found and
saved.
[17] Moreover, arcs EK (and) LM intersect between arcs LD
(and) DE. Therefore, the ratio of the sine of arc DL to the sine
of arc DK is compounded of the ratio of the sine of arc LM to the
sine of arc MH, and the ratio of the sine of arc EH to the sine of
arc EK. (But) each of the arcs LM (and) LD is equal to a quad-
rant, since L is the pole of circle DEZ which is the equator.
Therefore, the ratio of the sine of arc LD to the sine of arc DK
is compounded of the ratio of the sine of arc LD to the sine of
arc MH, and the ratio of the sine of arc HE to the sine of the arc
KE. Also, the ratio of the sine of arc LD to the sine of arc DK is
compounded of the ratio of the sine of arc LD to the sine of arc
MH, and the ratio of the sine of arc MH to the sine of arc DK.
So the ratio of the sine of arc MH to the sine of arc DK is as the
ratio of the sine of arc HE to the sine of arc EK. So, if we multi-
ply the sine of arc MH by the sine of arc EK, and divide by the
sine of arc EH, we will obtain the sine of the arc DK. If then we
find the arcsine (of this last quantity), it will be arc DK, which
is that which we designated the second arc. Now, arc MH is the
latitude of Mecca; arc EK is a quadrant, whose sine is sixty; and
arc EH is that whose sine we have saved. So, if multiply the sine
of the latitude of Mecca by sixty, and divide (the product) by the
sine which we saved, we will obtain the sine of arc DK. If then
we find the arcsine of (this last quantity), we will obtain arc DK,
which we have designated the second arc.
[18] Moreover, if the difference between the two longitudes was
less or more than a quadrant, or if it was a quadrant, then arcs
BT (and) EK intersect between arcs DL (and) AE at point H.
Therefore, the ratio of the sine of arc BA to the sine of arc AK is
compounded of the ratio of the sine of arc BT to the sine of arc
HT, and the ratio of the sine of arc HE to the sine of arc EK.
(But) each of BT and BA is a quadrant, since point B is the pole
of circle AG which is the horizon; therefore, the ratio of the sine
of arc BA to the sine of arc AK is compounded of the ratio of the
sine of arc BA to the sine of arc HT, and the ratio of the sine of
arc HE to the sine of arc EK. Thus, in a (manner) similar to
 IBN AL-HAYTHAM'S UNIVERSAL SOLUTION 159
what has preceded, the ratio of the sine of arc TH to the sine of
arc AK will be as the ratio of the sine of arc HE to the sine of arc
EK. So, if we multiply the sine of arc AK by the sine of arc EH,
and divide (the product) by the sine of arc EK, we obtain the
sine of arc HT. (But) arc AK is that whose sine is the auxiliary
sine - or it is the latitude of the locality if the difference between
the longitudes of the two localities was a quadrant; arc EH is
the one whose sine we saved - or it is the latitude of Mecca if the
difference between the two longitudes was a quadrant; and arc
EK is a quadrant, the sine of which is sixty. So, if we multiply
the auxiliary sine by the sine which we have saved - or the lati-
tude of the locality by the latitude of Mecca, this being the case
if the difference between the two longitudes is a quadrant - and
divide the product by sixty, then we obtain the sine of arc HT. If
then we find the arcsine of (this last quantity), we will obtain
arc HT. If we subtract it from ninety, the remainder will be arc
HB. So, arc HB is the one whose sine we have found and saved.
[19] Finally, and additionally, arcs BT (and) EK intersect
between arcs BA (and) AE at point H. Therefore, the ratio of the
sine of arc EA to the sine of arc AT is compounded of the ratio
of the sine of arc EK to the sine of arc KH, and the ratio of the
sine of arc HB to the sine of arc BT. (But) each of the arcs EK
and EA is a quadrant; so the ratio of the sine of arc EA to the
sine of arc AT is compounded of the ratio of the sine of arc EA
to the sine of arc KH, and the ratio of the sine of arc HB to the
sine of arc BT. So, if we multiply the sine of arc KH by the sine
of arc BT, and divide the product by the sine of arc HB, the
result will be the sine of arc AT. (But) arc KH is the arc which
we designated the first - or is the complement of the latitude of
Mecca; arc BT is a quadrant whose sine is sixty; and arc HB is
the last one whose sine we found and saved. So, if we multiply
the sine of the first arc - or the sine of the complement of the
latitude of Mecca, this being the case if the difference between
the two longitudes was a quadrant - by sixty, and divide by the
sine which we have saved last, then the result will be the sine of
arc AT. If then wefindthe arcsine of (this last quantity), we will
obtain arc AT, which is the azimuth arc. This is what we wanted
to clarify.
[20] Now if the latitude of the locality at which the azimuth is
required was northerly, and the difference between the two lon-
gitudes less than a quadrant, and if the arc which we designated
the second was equal to the latitude of the locality at which
160 AHMAD S. DALLAL

G-North G-North
Figure 1 Figure 2

T,H
A-North A-North
Figure 3 Figure 4

A-North
Figure 5
 IBN AL-HAYTHAM'S UNIVERSAL SOLUTION 161
G A

A-North G-North
Figure 6 Figure 7

G-North A-North

Figure 8 Figure 9

G-North G-North

Figure 10 Figure 11
162 AHMAD S. DALLAL

GrNorth A-North
Figure 12 Figure 13

A-North A-North
Figure 14 Figure 15

the azimuth is required, then the (corresponding) configuration

will be the first figure. The azimuth will then be the east-west
line, because arc AE is a quadrant, and arc BH is the latitude
circle which passes through the zeniths of the two localities.
[21] If the latitude was southerly, and the difference between
the two longitudes more than a quadrant, and if the second arc
was equal to the latitude, then the (corresponding) configura-
tion will be the second figure. The azimuth will then be the east-
west line, because arc AE is a quadrant, and arc DEHK is the
latitude circle, and because the difference between the two lon-
gitudes is DM, whose sine is equal to the sine of DM.
[22] If the latitude was northerly, and the difference between
the two longitudes greater than a quadrant, and if the second
arc was equal to the complement of the latitude, then the (cor-
responding) configuration will be the third figure. The azimuth
 IBN AL-HAYTHAM'S UNIVERSAL SOLUTION 163
arc will then be AT, which is the one we designated the first,
and it will be northerly.
[23] If the latitude was southerly, and the difference between
the two longitudes less than a quadrant, and if the second arc
was equal to the complement of the latitude, then the (corre-
sponding) configuration will be the fourth figure. The azimuth
arc will then be AT, which is the one we designated the first,
and it will be northerly.
[24] If the latitude of the locality at which the azimuth is
required was northerly, and the difference between the two lon-
gitudes greater than a quadrant, and if the second arc was not
equal to the complement of the latitude, then the (correspond-
ing) configuration will be the fifth figure. This is (true) because
the difference between the two longitudes is arc DM, whose sine
is the sine of arc DM; (moreover,) the latitude circle which
passes through the zenith, which is the upper point A, also
passes through the diametrically opposite lower point {nadir) B.
The azimuth arc will then be AT, and it will be northerly.
[25] If the latitude was northerly, and the difference between
the two longitudes was a quadrant, then the (corresponding)
configuration will be the sixth figure. The azimuth arc will then
be AT, and it will be northerly.
[26] If the latitude was northerly, and the difference between
the two longitudes less than a quadrant, and if the second arc
was greater than the latitude of the locality, then the (corre-
sponding) configuration will be the seventh figure. The azimuth
arc will then be AT, and it will be northerly.
[27] If the latitude was northerly, and the difference between
the two longitudes less than a quadrant, and if the second arc
was smaller than the latitude, then the (corresponding) configu-
ration will be the eighth figure. The azimuth arc will then be
AT, and it will be southerly.
[28] If the latitude was southerly, and the difference between
the two longitudes less than a quadrant, and if the second arc
was not equal to the complement of the latitude, then the (cor-
responding) configuration will be the ninth figure. The azimuth
arc will then be AT, and it will be northerly.
[29] If the latitude was southerly, and the difference between
the two longitudes was a quadrant, then the (corresponding)
configuration will be the tenth figure. The azimuth arc will then
be AT, and it will be northerly.
[30] If the latitude was southerly, and the difference between
164 AHMAD S. DALLAL
the two longitudes greater than a quadrant, and if the second
arc was greater than the latitude, then the (corresponding) con-
figuration will be the eleventh figure. (This is true) because DB
is equal to the latitude. The azimuth arc will then be AT, and it
will be northerly.
[31] If the latitude was southerly, and the difference between
the two longitudes greater than a quadrant, and if the second
arc was less than the latitude, then the (corresponding) configu-
ration will be the twelfth figure. The azimuth arc will then be
AT, and it will be southerly.
[32] If the locality was on the equator, and the difference
between the two longitudes was less than a quadrant, then the
(corresponding) configuration will be the thirteenth figure. The
azimuth arc will then be AT, and it will be northerly.
[33] If the locality was on the equator, and the difference
between the two longitudes was greater than a quadrant, then
the (corresponding) configuration will be the fourteenth figure.
The azimuth arc will then be AT, and it will be northerly.
[34] If the locality was on the equator, and the difference
between the two longitudes was a quadrant, then the (corre-
sponding) configuration will be the fifteenth figure. The
azimuth arc will then be AT", which is the complement of the lat-
itude of Mecca, and it will be northerly.
[35] As for arc AK, if the latitude was northerly, and the second
arc greater or smaller than the latitude of the locality, and if the
difference between the two longitudes was less than a quadrant,
then it is apparent from the seventh and eighth figures that the
sine of this arc, I mean arc AK, will be the sine of the comple-
ment of the latitude plus arc DK, which is the second arc: in the
eighth figure, arc AK is the sum of the complement of the lati-
tude, which is AD, and arc DK. In the seventh figure, arc GK is
the sum of the complement of the latitude, which is GD, and arc
DK, which is the second arc; the sine of their sum is the sine of
arc AK.
[36] If the latitude was northerly, and the difference between
the two longitudes greater than a quadrant, and if the second
arc was less or greater than the complement of the latitude of
the locality, then it is apparent from the fifth figure that: if the
second arc was less than the complement of the latitude, then
the sine of arc AK will be (equal to) the sine of the complement
of the latitude minus the second arc DK. If, (on the other hand,)
the second arc was greater than the complement of the latitude,
 IBN AL-HAYTHAM'S UNIVERSAL SOLUTION 165
then the sine of arc AK will be (equal to) the sine of arc DK
minus the complement of the latitude.
[37] If the latitude was southerly, and the difference between
the two longitudes greater than a quadrant, and if the second
arc was greater or smaller than the latitude, then it is apparent
from the eleventh and twelfth figures that the sine of the arc AK
will be the sine of the complement of the latitude plus arc DK,
which is the second (arc): in the twelfth figure, arc AK is the
sum of the complement of the latitude, which is AD, and arc
DK, which is the second (arc). In the eleventh figure, arc GK is
(the sum of) the complement of the latitude, which is GD, and
arc DK, which is the second arc; the sine of their sum, which is
(arc) GK, is the sine of AK.
[38] If the latitude was southerly, and the difference between
the two longitudes less than a quadrant, and if the second arc
was smaller or greater than the complement of the latitude,
then it is apparent from the ninth figure that: if the second arc,
which is DK, was less than the complement, then the sine of arc
AK will be (equal to) the sine of the complement of the latitude
minus the second arc DK. If, (on the other hand,) the second arc
was greater than the complement of the latitude, then the sine
of arc AK will be (equal to) the sine of arc DK minus the com-
plement of the latitude.
[39] If the locality was on the equator, and the difference
between the two longitudes was smaller or greater than a quad-
rant, then it is apparent from the thirteenth and fourteenth fig-
ures that arc AK will be what remains after subtracting arc DK,
which is the second (arc), from ninety.
[40] It is thus apparent, that the arc which we found by means
of the above mentioned calculation is the azimuth arc. In each
of the (above) configurations, this (arc) will be arc AT.
Moreover, We have found the sine of this arc in all (the above)
diagrams. It is apparent that if the locality was east of the ka'ba,
then the arc will be west of the locality, and if (the locality) was
west of the ka'ba, then the arc will be east of it. It is also appar-
ent that if, on the circle which falls in the plane of the horizon,
we mark away from the meridian line an arc equal to this
(azimuth) arc, in the direction which was established in the fig-
ures, and if we connect by a straight line the end (mark of this
arc) and the center of the circle, which is the locality at which
the azimuth is required, then this line will be the intersection
between the plane of the horizon and the altitude circle which
166 AHMAD S. DALLAL
passes through the zeniths of the two localities - I mean the
zenith of the ka'ba, and the zenith of the locality at which the
azimuth is required. This will be the line which demarcates the
azimuth of the qibla.
[41] It is therefore apparent that: if we found the azimuth arc
by means of the above mentioned calculation; if we adjust a
plane parallel to the horizon, draw a circle in it, and construct
in it the meridian line which passes through its center; if we
mark on the circumference of the circle an arc equal to the arc
which we found by means of calculation, in the direction which
was established in the figure; and if we connect by a straight
line the end (mark of this arc) and the center of the circle, then
this line will be the one that demarcates the azimuth of the
qibla. This is what we intended to clarify.
[42] It is also possible to extract the azimuth by using an
instrument. This can be (done) by (using) an armillary sphere
(dhdt al-halaq), if it has a ring that corresponds to the latitude
circle, and another ring which corresponds to the horizon (cir-
cle). (First,) we move the pole of the armillary sphere above its
horizon through an amount equal to the altitude of the pole in
the locality where the azimuth is required. We then rotate one
of the rings of the (armillary) circle around the pole of the equa-
tor until it reaches the point of the equator whose distance from
the second meridian ring is equal to the (difference in) degrees
between the two longitudes - 1 mean the longitude of Mecca and
the longitude of the locality at which the azimuth is required.
Next, we rotate the ring that corresponds to the latitude circle
until it reaches the point on the ring around the pole of the
equator whose distance from the equator circle is (equal to) the
latitude of Mecca. We mark the point where this ring, (that
corresponds to the latitude circle,) intersects the ring that cor-
responds to the horizon. We then determine the arc between
this point and the meridian circle, and the direction of this arc.
This arc will be the azimuth arc for the locality at which the
azimuth is required. This is what we wanted to clarify.
[43] The essay on (finding) the azimuth of the qibla is com-
pleted. Praise be to God, the Lord of all creations.

MATHEMATICAL COMMENTARY

The problem of determining the direction of the qibla from the

terrestrial coordinates of any locality and of Mecca is a problem
 IBN AL-HAYTHAM'S UNIVERSAL SOLUTION 167
of spherical geography. This problem, however, was treated by
Ibn al-Haytham, as by many other medieval astronomers, as a
problem of spherical astronomy. The equivalent astronomical
problem is to consider the zeniths of the locality and of Mecca,
and to determine the azimuth of Mecca measured along the
horizon circle. The azimuth is usually measured from the south
point, although, when it is more convenient, it may be measured
from the north point.
Ibn al-Haytham starts by denning the azimuth of the qibla as
the intersection between the local horizon and the great circle
passing through the zeniths of Mecca and the locality (§ [3]).
Ibn al-Haytham then gives a summary of the different ways by
which he proposes to find this azimuth. He first lists five special
cases in which both zeniths fall on the same meridian circle (§
[4]). He then considers the general case, and lists all possibili-
ties that may arise, as well as the value of the azimuth arc and
its direction in each case. He also indicates how to determine
whether the arc is easterly or westerly (§ [5]-[14]). I propose
now to survey these procedures using modern mathematical
symbols to represent what Ibn al-Haytham describes completely
in words, as was standard in medieval scientific writing.
Denote the latitude of the locality by (p, that of Mecca by <pm,
and the difference between the longitudes of the two by AL. Ibn
al-Haytham defines the following arcs:
The first arc = a = arcSin —
R
in m
The second arc =B = arcSin ^ _—
Since
where cpm is the complement of <pm, and a is the complement of
a. Moreover, Sin a = R sin a, where R is the radius of the base
circle for which the sine is calculated, and is equal to 60.33
Finally, The auxiliary sine = Sin y where y is a function of <p
and a (see below).
Denoting the longitude of the locality by L, that of Mecca by
Lm, and the azimuth arc by q, the first part of Ibn al-Haytham's
essay (§ [4]-[14]) can be summarized in the following tables:
33
On the use of base 60 see Kennedy, "The history of trigonometry," pp. 3-29; also
on the use of R see Otto Neugebauer, A History of Ancient Mathematical Astronomy,
3 vols. (New York, 1975), Part 3, pp. 1115-6.
168 AHMAD S. DALLAL
a. Special cases: This table gives the direction of the azimuth of
the qibla (column 6) as a function of the position of the locality
relative to Mecca.
Case <P A<p AL Shortest Distance Between Azimuth of the Qibla
the Locality and Mecca

1 North of *0 0 Meridian Line - Southerly

ka'ba
2 South of *0 0 Meridian Line - Northerly
ka'ba
3 South of 180° 180° Any Direction
Equator
4 South of <180° 180° Passes through the North North part of the Meridian
Equator Pole Line
5 South of > 180° 180° Passes through the South South part of the Meridian
Equator Pole Line

b. General Case: In this table, each case is denned by the first

three entries in a row. The fifth column indicates the direction
of the azimuth arc, and the sixth column gives its value. Since
negative values were not used, column 4 indicates the absolute
value of arc AK = y which should be substituted in each case in
the equations of column 6.
<p AL r Direction of the 9
Azimuth Arc jj_

Northerly ALOW P= 9 E-W Line: Locality East of Mecca => q to the West

Locality West of Mecca => q to the East

Southerly 90°<^<180° P = V E-W Line: Locality West of Mecca => q to the East

Locality East of Mecca => q to the West

Northerly 90°<^<180° P=9 Northerly q =a

Southerly 4L<90° P= 9 Northerly q= a

Northerly 90°<4L<180° P<9 9-P Northerly

P> 9 P-9 Northerly

Southerly P><P 9+P Northerly

Sina x60
p<tp 9+P Southerly

Northerly AL<90° P><p <p+P Northerly

p<<p 9+P Southerly

Southerly 4L<90° p< ~y 9-P Northerly
p> ~q> P-9 Northerly

North or AL-90"
Sm(90 - arcSin '-)
South ou

On Equato AL=90° i = 9n
 IBN AL-HAYTHAM'S UNIVERSAL SOLUTION 169
c. East and west directions of the azimuth of the qibla:

Point From Which the L vs. Lm f Azimuth Arc

Longitude is Measured

West Point L>Lm East of Mecca Westerly

East Point L<Lm East of Mecca Westerly

East Point L>Lm West of Mecca Easterly

West Point L<La West of Mecca Easterly

After this introductory summary, Ibn al-Haytham describes the

procedure for finding the actual direction to face, once the
azimuth arc q has been determined by calculation (§ [15]). This
is done simply by projecting the results of calculation on a plane
which is leveled parallel to the horizon. Ibn al-Haytham then
proceeds to prove his propositions (§ [16]-[19]). In this proof,
Ibn al-Haytham uses the "theorem of six quantities," which is
the name given by Arab astronomers to the Menelaos theo-
rem.34 No three dimensional diagram is provided in the manu-
scripts for the proof of the general case, and Figure A has been
constructed based on the text. This figure shows the celestial
sphere as seen in the locality for which the azimuth is required.
The great circle EAZG is the local horizon, and its pole B is the
zenith of the locality. Circle ADKBLG is the local meridian.
Point L is the north pole and circle EMDZ is the celestial equa-
tor. Point H is the zenith of Mecca, and circle LHM is the merid-
ian of Mecca. The great circle passing through E, H and Z inter-
sects the local Meridian at K. Finally, circle BUT is the altitude
circle of Mecca which passes through the zeniths of the locality
and of Mecca, and which is perpendicular to the local horizon.
In this diagram, arc LG = arc BD = <p, which is the latitude of
the locality, and arc LB = arc DA = <pj arc HM = cpm, which is
the latitude of Mecca, and arc LH = q>m; arc DM is the differ-
ence AL between the longitude of Mecca and that of the locality.
Finally, arc AT (= q) is the azimuth arc which defines the
required direction of Mecca on the horizon plane.
Following is the proof exactly as Ibn al-Haytham presents
it, but written using modern mathematical notation. Remarks
in square brackets are not in the original but were added for
clarity.
34
On the Menelaos theorem see Neugebauer, A History of Ancient Mathematical
Astronomy, Part 1, pp. 26-9; also see Kennedy, "The history of trigonometry," pp. 12-15.
168 AHMAD S. DALLAL
a. Special cases: This table gives the direction of the azimuth of
the qibla (column 6) as a function of the position of the locality
relative to Mecca.
Case f Aq> AL Shortest Distance Between Azimuth of the Qibla
the Locality and Mecca

1 North of *0 0 Meridian Line - Southerly

ka'ba
2 South of *0 0 Meridian Lane - Northerly
ka'ba
3 South of 180° 180° Any Direction
Equator
4 South of <180° 180° Passes through the North North part of the Meridian
Equator Pole Line
5 South of >180° 180° Passes through the South South part of the Meridian
Equator Pole Line

b. General Case: In this table, each case is defined by the first

three entries in a row. The fifth column indicates the direction
of the azimuth arc, and the sixth column gives its value. Since
negative values were not used, column 4 indicates the absolute
value of arc AK = y which should be substituted in each case in
the equations of column 6.
AL , - , r Direction of the •
Azimuth Arc q

Northerly 4L<90° P=<P E-W line: Locality East of Mecca => q to the West

Locality West of Mecca => q to the East

Southerly 90°<4L<180° P=V E-W line: Locality West of Mecca =» q to the East

Locality East of Mecca => 17 to the West

Northerly 90°<4Z,<180° P= V Northerly q=a

Southerly 4L<90° P= V Northerly q=a

Northerly 90°<4L<180° P<V V-P Northerly

P> V P-V Northerly

Southerly p><p 9+P Northerly

_. r Since x 60 .
p<<p 9+P Southerly n • ,n,i o- Since xSiny^
Northerly 4L<90° p>(f> V+P Northerly

p<<f> V+P Southerly

Southerly 4L<90° p< <p (P-P Northerly
P> V P-9 Northerly
Sin<pnx6Q
North or AL--S0*
?• too s' S'n(PxS'ny\
South 60

On Equato AL=90°
 IBN AL-HAYTHAM'S UNIVERSAL SOLUTION 169
c. East and west directions of the azimuth of the qibla:

Point From Which the L vs. Lm <P Azimuth Arc

Longitude is Measured

West Point L >Lm East of Mecca Westerly

East Point L <Lm East of Mecca Westerly

East Point L>Lm West of Mecca Easterly

West Point L<Lm West of Mecca Easterly

After this introductory summary, Ibn al-Haytham describes the

procedure for finding the actual direction to face, once the
azimuth arc q has been determined by calculation (§ [15]). This
is done simply by projecting the results of calculation on a plane
which is leveled parallel to the horizon. Ibn al-Haytham then
proceeds to prove his propositions (§ [16]-[19]). In this proof,
Ibn al-Haytham uses the "theorem of six quantities," which is
the name
34
given by Arab astronomers to the Menelaos theo-
rem. No three dimensional diagram is provided in the manu-
scripts for the proof of the general case, and Figure A has been
constructed based on the text. This figure shows the celestial
sphere as seen in the locality for which the azimuth is required.
The great circle EAZG is the local horizon, and its pole B is the
zenith of the locality. Circle ADKBLG is the local meridian.
Point L is the north pole and circle EMDZ is the celestial equa-
tor. Point H is the zenith of Mecca, and circle LHM is the merid-
ian of Mecca. The great circle passing through E, H and Z inter-
sects the local Meridian at K. Finally, circle BUT is the altitude
circle of Mecca which passes through the zeniths of the locality
and of Mecca, and which is perpendicular to the local horizon.
In this diagram, arc LG = arc BD = <p, which is the latitude of
the locality, and arc LB = arc DA = <p; arc HM = q>m, which is
the latitude of Mecca, and arc LH = q>m; arc DM is the differ-
ence AL between the longitude of Mecca and that of the locality.
Finally, arc AT (= q) is the azimuth arc which defines the
required direction of Mecca on the horizon plane.
Following is the proof exactly as Ibn al-Haytham presents
it, but written using modern mathematical notation. Remarks
in square brackets are not in the original but were added for
clarity.
34
On the Menelaos theorem see Neugebauer, A History of Ancient Mathematical
Astronomy, Part 1, pp. 26-9; also see Kennedy, "The history of trigonometry," pp. 12-15.
170 AHMAD S. DALLAL
1) Applying the Menelaos theorem to the spherical triangle
HEM, with DEL as the transversal of this triangle we obtain:
SinED SinEK SinHL
x
SinDM SinKH SinLM
but
EK = ED = 90°
therefore,
SinED SinED SinHL
__ v^

SinDM ~ SinKH SinLM'

and
SinED _ SinED SinKH
SinDM ~ SinKH SinDM''
therefore,
SinKH _ SinHL
SinDM ~ SinLM'
SinDM x SinHL
SinKH =
SinLM
and SinDM x SinHL
KH = arcSin
SinLM
but
DM = AL, = ffl , &LM = 90°,
therefore,
„„ _. SinALxSinq>Ymm
KH = arcSin—
60

[which is defined by Ibn al-Haytham as the first arc a; also

EH = 90° - KH, therefore, EH = a.]
2) Applying the Menelaos theorem to the spherical triangle
LKH, with DME as the transversal of this triangle we obtain:

SinDL SinLM SinEH

SinDK SinMH SinEK
but
LM = LD = 90°,
therefore,
SinLD SinLD SinHE
v
SinDK SinMH SinKE
and
 IBN AL-HAYTHAM'S UNIVERSAL SOLUTION 171
SinLD SinLD SinMH
SinDK SinMH x- SinDK
therefore,
SinHE _ SinMH
SinKE ~ SinDK'
SinMH x SinKE
SinDK =
SinEH
and
_. SinMH x SinKE _. Sinmx 60
DK = arcSin : : = arcSin——
SinEH Sina

[which is defined by Ibn al-Haytham as the second arc /J; also

AK = q> + DK, which is denned by Ibn al-Haytham as the auxil-
iary sine y.]
3) Applying the Menelaos theorem to the spherical triangle
BKH, with ETA as the transversal of this triangle we obtain:
SinAB _ SinBT SinHE
SinAK ~ SinHT SinEK'
but
BT = BA = 90°,
therefore, SinBA SinBA SinHE
SinAK SinHT SinEK
and
SinBA SinBA SinTH
SinAK SinTH SinAK '
therefore,
SinHE SinTH
SinEK SinAK'
and
c. irrr SinAKxSinEH
SinHT = ,
but SinEK
therefore, SinAK = Siny, EH = a, and EK = 90°

irr c.
Siny x Sina
and HT = arcSin = h,
60
BH = 90°-HT = ~h~.
172 AHMAD S. DALLAL
In the special case when AL = 90°, point K will coincide with
point L, and arc AK will be equal to 180° - cp, or Sin AK - Sin
(p. Also arcs KHE and LHM will coincide, therefore, arc EH will
be equal to arc MH, which is the latitude q>m of Mecca. Thus in
this case we will have:

60
4) Applying the Menelaos theorem to the spherical triangle
ETH, with BKA as the transversal of this triangle we obtain:
SinEA SinEK SinHB
— w

SinAT SinKH SinBT'

but
EK = EA = 90°,
therefore,
SinEA _ SinEA SinHB
SinAT ~ SinKH SinBT'
and
SinEA _ SinEA ^ SinKH
SinAT ~ SinKH X SinAT '
SinKH SinHB
SinAT SinBT
0X1
Arr _. SinKH x SinBT _. 5inax60
AT = q = arcSm = arcSin =—.
SinHB Sinh
In the special case where AL = 90°, KH will be equal to the com-
plement of the latitude of Mecca, and thus the final equation
will be:
Sm<pmx60
Zjn
q = AT = arcSin -= .
Sinh
Finally, if in the last two equations we substitute the values of
h and a which were calculated in steps 1 through 3, we will get
the exact values listed in table (b) above.

Construction of the Figures:

After calculating q, Ibn al-Haytham constructs fifteen separate
 IBN AL-HAYTHAM'S UNIVERSAL SOLUTION 173
diagrams for all the possible positions of the locality with respect
to Mecca, in order to determine the direction of the azimuth arc
(§ [20]-[40]). The figures are completely missing from the Cairo
manuscript, and are not legible in the Berlin one. Therefore the
figures had to be reconstructed following the description pro-
vided in the text. In the following table, the figures mentioned in
the text are rearranged according to the cases rather than the
figure numbers assigned by Ibn al-Haytham. Columns 2, 3 and 4
define each case, and columns 5 (corresponding to column 5 in
table b) and 6 (corresponding to column 4 in table b) are deduced
from the constructed diagrams.

Figure # Locality B direction of g SinAK

1 Northerly <90° 0= V East-West Line

7 Northerly <90° B> V = AT, Northerly Sin(9+B)

8 Northerly <90° B< V = AT, Southerly Sin(9 + B)

3 Northerly >90° B = <P = o, Northerly

5 Northerly >90° B * <P = AT, Northerly B< 9=*Sin(.9- B)

B > P=» Sin 03 - 9)

6 Northerly = 90° = AT, Northerly

2 Southerly >90° B = <t> East-West Line

11 Southerly >90° B>q> = AT, Northerly Sin(<P+ 0)

12 Southerly >90° B<9 = AT, Southerly Sin(9+ B)

4 Southerly <90° / 3 = <P = a, Northerly

9 Southerly <90° B* 9 = AT, Northerly B< <P=>Sin(<P- B)

B> 9=>SinW- <P)

10 Southerly = 90° = AT. Northerly

13 On Equator <90° = AT, Northerly P

14 On Equator >90° = AT, Northerly P
15 On Equator = 90° = AT= <P~, Northerly

To determine the direction of the azimuth arc of the qibla, Ibn

al-Haytham projects the celestial sphere onto the plane of the
local horizon, and finds the position of Mecca relative to the
locality on this plane. To reconstruct the diagrams I first con-
structed the three dimensional diagrams, and then projected
174 AHMAD S. DALLAL
them using approximate mapping on the horizon planes for each
of the given cases. The three dimensional figure presented here
as figure A corresponds to case number 8. Figure 8 is obtained
from figure A by projecting the different circles of the sphere on
the plane of the local horizon EAZG with the zenith B of the
locality as the pole of projection. In figure 8, the local meridian is
perpendicular to the plane of projection and coincides with its
vertical diameter ABG; the horizontal diameter EBZ is the east-
west line; the north pole L falls in the north part of the circle,
which, in accordance with medieval convention, is maintained at
the lower side of the circle. EMDZ is the projection of the equa-
tor, and EHKZ is the projection of the great circle through E, H
and Z. DM marks the projection of the arc on the equator which
is equal to AL, and LHM is the projection of the meridian of
Mecca. LM intersects EKZ at H; therefore, point H is the projec-
tion of the zenith of Mecca on the local horizon, and line BHT is
the intersection between the local horizon and latitude circle
passing through the zeniths of Mecca and the locality. This line
defines the direction of the azimuth of the qibla, while arc AT,
which was calculated in the previous section, defines the value
of this arc. It is apparent that in this case, the azimuth of the
qibla is southerly. Also in figure 8 GL = BD = q>; LB = AD = <p;
DM = AL; KH = a; HM = q>m; LH = <pm; and DK = p.
Now, to illustrate the construction of the other figures, con-
sider figure 1 which represents a special case of figure 8. In case
1, as in case 8, the locality is northerly, and AL is less than a
quadrant. However, in case 1 the second arc P is equal to the lat-
itude q> of the locality. Therefore, DK = BD, and points K and B
will coincide. The projection of circle EHKZ will thus coincide
with the straight line EBZ, and the azimuth arc will fall along
the east-west line.
Similarly, the other cases can be constructed by projecting
each three dimensional configuration into the plane of the cor-
responding local horizon. Depending on the value of AL, and on
the northerly or southerly position of the pole of projection B,
the north point on the local horizon will fall on either point A or
point G. Thus, different characters were used to refer to the
north point in the figures, but it consistently falls in the lower
side of the circle. E always refers to the east point, while Z
refers to the west. The positions of E and Z, however, switch
from the right side to the left side of the figures depending on
the part of the sphere where the pole of projection B falls.
 IBN AL-HAYTHAM'S UNIVERSAL SOLUTION 175
Paragraph [41] (like § [15]) is on the actual determination of the
direction of the qibla in a specific quadrant, once the value of
the arc and its direction have been theoretically determined.
Ibn al-Haytham then presents a second method for finding the
direction of the qibla by using the instrument known as the
armillary sphere which represents the problem in a three-
dimensional framework (§ [42]). The azimuth arc is found by
aligning the different rings of the instrument so that they cor-
respond to their celestial counterparts. The azimuth arc can
then be read off directly, without any calculation whatsoever.35 *
36
On the use of armillary sphere see E. S. Kennedy, "Al-Kashi's treatise on astro-
nomical observational instruments," in Kennedy et al, Studies in the Islamic Exact
Sciences, pp. 394-404.
The author would like to express his gratitude to Professor F. Sezgin for supply-
ing related material which was of use during the study, and to Professors E. S.
Kennedy and G. Saliba, who read this paper and made valuable comments on it.
Special thanks go to Professor D. A. King who familiarized me with the topic in the
first place, supplied copies of the Cairo and Berlin manuscripts, and guided me
through the study. The author remains, however, solely responsible for any errors
that may appear in this paper.
176 AHMAD S. DALLAL

x]
^ [t]
I IJ ^

JIMJ j . <L«.<Tll Jl«L-i,l ilaiJL i

^ll j i t ^ J ^ i l l J^LlI ^k «iJJI ^i7.

[t]

iiirt*! i a i - ^

.til jjii lU "SjJ*" L

fSs*« ~ Jy
 IBN AL-HAYTHAM'S UNIVERSAL SOLUTION 177

Ob -j-i^l -»° "

.-LiHj 0

J t * ^ <f>

jjj

U J-iij i£giillM j

OH

\o d\ '

^ *^->-^ J—»»-
Ju L4
178 AHMAD S. DALLAL

1)1 (j-,211 »1A Culi" j j j [v]

Ui . ^ i ^ l ^12 J* Ul&ll o-^ill Ju>i bli .S^b ^ij ^ Jit C&J>^I CW

 IBN AL-HAYTHAM'S UNIVERSAL SOLUTION 179

La l.:..,;.^

ULU \s6 iSJi\ <rJL\ J jiUadI v^J-l . . ^

^* «• • •» •

'UL-i ^ ^ l j l i ' j r»ji\i gj ^ ^ J l OH ^ LP* J^" ob [ u ]

L i <!^4 \j>j£- v^r?" tr* ^"-••'•"•^ *-^ > -r'>^ ^ f-*

>. L i < jVI «LLi>. (5 jjl

180 AHMAD S. DALLAL
4-i V ^LUI ^ l ^ l l . j i ^ j l ^ j,U '

^-U jli" j l i '^>JaJJ >j -i-JL^i >

-.,«.,..II ij^yii i*•.,»> It

l i i 4,..«.VJl

\o ^ ' l i j i U J^kll j l ^ j .iSU J^U ,>« y i i «-i^i>l J^L j l ^ lit jjSlj dljj

>. {f 'li^i-U JjiJIj . j-ij-U J^i. ,>. j i i i i£« J>J» jlS" j l . j j

I ^^^ f ^
 IBN AL-HAYTHAM'S UNIVERSAL SOLUTION 181
jljJLU U j j J *\j>-\ jL^-!l t.iirt'i Jai- j j J ^ l^-U J.aa'ij . <JJL..T.«

Jo /

i ULJ.(j « j

a I.U1-.I

• j*

I *j IJ > > L_> I t

JU—J ^2 ^dl JLu-JI S>jl J l , ^ l l l j

\0 *>Jaf- »^jb |»*>3j / .>"-••.-II ^>-ji I» I -» • J» r- v \o

»il j JJ *JiU . til * • » » j j b ^ j i *• i •

' r

wj Ji" ,jX3. * J

I *j

i T(V)-V.- |j].j..3N - (j.

182 AHMAD S. DALLAL
. Mr (j"i yJ ^ ijrjr /ir . t/\7 ^ U^J" J ,\!. (,f> *

^ t r y v * # **—j i £

^ V—^ ^ J try v*e- Jl •£_ ^

•J
• r ^ t-^y ^"^ ^ ?->*•' C

II,-^- J$\ <y» r • cr>»» • r *

4 ...•••> i . J i dl » L,y* 4 J . J J (j—y" CtH W * ^i»Uu Ji <J^* U i j l j [ ^ V]

t^l ^ <J tr>* V*rf- *i— i tV *-*J>« ^ •s t r y v-^rt- t^l J J

t r y v^rt-

^ t r y "r-e- J l • c °"^ <r*s?' ^^^ < ^ ' C C

J
•**•> i_^[ t) t r y V*s^ *:""*'j • *

* ^ tj*ij* \+*f- *j •"*s »<3 i j ^ y >';•*»•*• (jJ; T" ^ t r y v*s?

 IBN AL-HAYTHAM'S UNIVERSAL SOLUTION 183
»\\ i . J » ^-ji y - » . ^ £ ^ ^ y v-e- LiJj-^ ••i^i . tsl • u-jS

^ • tryj •**-• <>^ t** c r ^ - ^

u 9-j^. ><Uai>- US' / tsJJi v ^ ' ^

»< i . (_. Lwjt • I . J A

-»•;»• <»"'O ( j ^ *-«J^« o I (_>"_jJ <-«;>iw _ ] l I i_i

I«

* t>-y vs*- J l • ^ t r y *-# *->—j (>y • J» ^ u-y ^rs?- J l ' v

try V*e- ^ ^»- ^ * o*£ "r-s?- J * «^«-'"«j r »

-* O1^ ol

- [o)j•tv^- IJHJ
184 AHMAD S. DALLAL

UJ J J AJ^I < t o j l j >'ljs>-i [^

'* 'T

I«

I« I

^ - t5JJI j i i > u r tf">* '.y>

H* • i
y*—*

^ jlf Ity [^ •] / /
«^j ^-« J i l ^ y a J I cW

JSLiJLi <C
 IBN AL-HAYTHAM'S UNIVERSAL SOLUTION 185
1 1
186 AHMAD S. DALLAL

AJiJ,
 IBN AL-HAYTHAM'S UNIVERSAL SOLUTION 187

\r
188 AHMAD S. DALLAL

llj .-Lilsll S j ^ l y i JSLUli , ji^JJ <UjL~. JLilsll

,> i V^

Ob
JSLiJU . ^ ^ J ! ^UxJ *^L— JJUH u->iJI i i V i j

<i> I >r

j^" ub M

[^] JUL-j | V ] jS-U ^» TO J^ull \ N-NA -

IBN AL-HAYTHAM'S UNIVERSAL SOLUTION 189

.iJU-i ^ j i\r I u.,..II

> i» I <"«.«. nil

o^ Jb [v • ]

T t l i - t f - II r • f* - I t I . -N!

» .1 II - tl IS II • II l-f - I II >ll

o L. J j J , . jLfJI Jo« C ^ ^ j l l ol^ ub [ rT ]

i]» I >"......11

. •jL- ^T -
190 AHMAD S. DALLAL

fef M i l l ^yS\ Cjtf, /UL-i j t f j l ,>^JI 0^4 . id t ^jS UU [To]

o <>• ^ CsV*11 CW *- cPij / U U - i J>

» u^j«W?- o^^1 0^ Ob l

j ' d j >• - [ 3 ] J j > J J N - [ 3 ] J j <JJV

 IBN AL-HAYTHAM'S UNIVERSAL SOLUTION 191

' l H l a £ 0 O- ^ C f i J ^ 1 OH

I »L
»l£ <U UA^LU d)

^ j l *J' o

JI ^k i J - l dUi
192 AHMAD S. DALLAL

<» \'*\ bl l ^w JJ

(j o>- L4-J L*-jj«J^

JjjlJJI OH

i j .XJ^I j ! r\j>£—\

4_J

,>« «JULJ (iJJI

(i jL^-JI J

V4•,A*t ^

. *iLi lijj
IBN AL-HAYTHAM'S UNIVERSAL SOLUTION 193

.a. & jsi"