Geographical Coordinates Of Localities From Islamic Studies

By E. S. & M. H. Kennedy

This article is extracted from Fuat Sezgin’s book: Jubilaumsband zum dreiBigjahrigen Bestehen des Institutes fur Geschichte der Arabisch-Islamischen Wissenschaften In europaischen Sprachen erschienene Vorworte zu Publikationen des Institutes aus den Jahren 1984 bis 2011, Institut fur Geschichte der Arabisch-Islamischen Wissenschaften an der Johann Wolfgang Goethe-Universitat, Frankfurt am Main, 2011, pp. 191-199.


The list of place names, with latitudes and longitudes, which is here presented, has been assembled from some seventy-four sources which range in size from 656 to two localities only. For the most part, these lists have been found in astronomical handbooks, zijes, where they enabled the astronomer to reduce observations reported from one locality to positions consonant with another. Some of the largest collections were intended to serve as a basis for plotting a world-map; others are found in works of descriptive geography. Isolated tables, frequently anonymous, have been encountered in bound volumes of manuscript treatises on various subjects. Finally, astronomical instruments, principally astrolabes, sometimes have geographical lists engraved upon them. Some have been incorporated into this collection.

Although the center of interest has been the medieval Islamic material, we have not hesitated to exploit sources which are earlier or later than the time span thus implied, provided that these are relevant to the main objective. Hence, coordinates due to Ptolemy (d. c. 180 A.D.), influential among Muslim geographers, will be found. At the other end of the time scale, some very late lists have been included because the entries seem to stem from much earlier documents.

In general, each entry in the list includes a place name, in Latin characters and in Arabic, longitude, latitude, code designations of the source, location within the source, remarks, and cross references. Each of these categories is commented upon below.

Place Names in Latin Characters

Quite frequently, a particular city will have two or more designations. Thus modern Istanbul is also known as Constantinople, and is most frequently found in the texts as Qusṭanṭīnīya, occasionally as Bīzanṭīya. But to insure that all occurrences of a particular locality can easily be retrieved from the computer storage for statistical purposes, all have been stored and printed under one name, chosen more or less arbitrarily. All other designations of the place known to us have been inserted in the list in the order of the Latin alphabet, whence cross references direct the reader to the single name under which all the coordinates appear.

A Q after a place name indicates that the reading is doubtful. Occasionally a single source gives the same place name twice, with different coordinates. Perhaps the original compiler conflated two or more different lists and carelessly failed to eliminate a duplication of cities. Under such circumstances, a 2 is entered behind the place name having the less customary coordinates. If for this city other sources give the same coordinates, 2’s are entered with them also.

Place Names in Arabic

Most of the sources are unpublished manuscripts written in Arabic characters by scribes who were unfamiliar with many of the names they were copying. It is not surprising that words, particularly of Greek, Persian, Turkish, or other provenance, should become grotesquely corrupted in the course of successive recopyings over the centuries. In most cases experience, and comparison of the coordinates with those of other localities, make it possible to restore the original form. There are names, however, which we are unable to read, and the entries which appear are the products of subjective guesswork.

The commonest source of difficulty in reading place names stems from the role played by dots in distinguishing between characters. A letter appearing at the beginning or in the middle of a word is nūn, tāᵓ, thāᵓ, bāᵓ, yāᵓ, or pāᵓ, depending only on the number of dots, between one and three, placed above or below the line. It is very easy to omit the dots, or to put them down wrongly, or to combine them wrongly with the dots of contiguous letters.
Generally speaking, vowel sounds are not indicated in the sources. The reader is expected to know beforehand the vowels in the word at hand.

The only distinction between a terminal hāᵓ and a tāᵓ marbūṭa is the absence or presence of two dots. Many of the sources are in Persian, where the two letters are commonly confused, and the scribes tended to omit the dots entirely. In like manner, there being no definite article in Persian, the scribes exhibited a cavalier attitude toward the omission (or inclusion) of the Arabic al-.

Other difficulties are described in the section on longitudes and latitudes.

In order to facilitate the eventual identification of doubtful names, an effort has been made to enable a user who reads Arabic to see each name as it appears in the source. To this end, a code was adopted, hampered by the fact that when the collection was commenced, the only symbols accepted by the computer then available, aside from a few punctuation marks, were the upper case letters of the Latin alphabet, plus the ten decimal digits. In this publication, the coded names have been turned back into Arabic, with dots missing or misplaced as in the source. For any locality in the alphabetical listing for which the space in the Arabic column is blank, the Arabic form is taken to be the same as the Arabic name appearing above it. For names obtained from Greek or Latin sources these spaces are, of course, blank.

Longitudes and Latitudes

Coordinates are reported in degrees and minutes of arc. Where the location of a city is well-known and modern coordinates are available they have been entered in the list. For purposes of computation, in order to appear as positive numbers, modern longitudes west of Greenwich have been recorded as east of Greenwich. Thus a longitude of 10° west of Greenwich is entered as 350°.

When a source clearly indicates that a particular latitude is south of the equator, a minus sign may have been entered before the degrees number.
Coordinates in the source are given as combinations of Arabic letters in the abjad numeral system, whereby integers are in a non-place-value decimal system, whereas minutes are in sexagesimals. There are well established conventions whereby, for instance, the two dots under the line to indicate a yāᵓ (=10) are unambiguously omitted, whereas the single dot over the line denoting a nūn ( = 50) must be inserted. Thus (yāᵓ, ṭāᵓ) is 19, whereas (nūn, ṭāᵓ) is 59. A careless scribe need only omit a dot, or insert one which should not be there, to introduce an error in the coordinates. Zāᵓ (=7), over which the dot is omitted, is easily misread as dāl (=4), or wāw (=6), and conversely. Jīm (=3) customarily has its dot omitted, and to distinguish it from ḥāᵓ (=8), its tail is left off, the tail of the ḥāᵓ being retained. When occurring at the beginning of a number, the only difference between fāᵓ (=80) and qāf (=100) is in their dots, fāᵓ having one and qāf two. It is easy for a scribe to get the dots wrong or to leave them out entirely. Where common sense or a comparison with other sources clearly indicate it, we have not hesitated to read in a missing dot, or to read a zāᵓ as dāl, and so on. Otherwise, we read the text as it is, making a plausible guess if necessary.

Zero Meridians

Unlike latitudes, which are universally measured from a natural reference, the terrestrial equator, longitudes are reckoned from a more or less arbitrarily chosen fixed meridian. Ptolemy measured longitudes east from the Fortunate Isles (Ar. al-Jazāᵓir al-Khālidāt), the Canaries, and about half our sources follow him. In the description of sources, the code designations of these tables are followed by a C. For all the other sources except three, corresponding longitudes tend to be ten degrees less than those of the C category, and some of the authors state explicitly that their base meridian is the “western shore of the encompassing sea (Atlantic)” (sāḥil al-baḥr al-muḥīṭ al-gharbī), ten degrees east of the other. Such tables are designated by an A. It is not clear how this distinction came about, for Nallino has shown (p. 490 of NAL in the list of sources) that it was not the intention of the early geographers of the Ma°mun school to shift the base meridian.

Two sources, HAM and HAB, measure longitudes west from a base meridian in the Far East. The author of the first says, in effect, that the relation connecting his longitudes with those of Ptolemy is L + Lp = 193.5°, where the subscript P denotes Ptolemy. Finally, one source, UTT, records longitudes from Basra (in Iraq) presumably his station.

Source Codes

To each of the sources a triliteral code has been assigned. The detailed description of sources which follows has been arranged in alphabetical order of the source code. In some cases, coordinates have been found in a secondary source with attribution to a primary source. For these, both codes are displayed. Thus, an entry found in Bīrūnī’s Masudic Canon is designated BIR. But a place whose coordinates are reported by Abū’l-Fidāᵓ as originating with Bīrūnī is designated FID BIR. For ease of computer operations a three digit numerical code has also been assigned, the numbers being, to the extent possible, in chronological order of the source.


This column gives for each source the folio or page, column, and location within the column where the particular entry is found. Peculiarities of individual sources are noted in the description of sources below. Short titles in boldface are references to the bibliography which immediately follows the tables.


Other than for cross references, this column is used to give a name form in the source which differs from that given under the place name column, or to indicate other variant reading possibilities. Lack of space in this and the reference column has occasioned the use of unorthodox abbreviations and cryptic remarks. The notion is that incomplete information inelegantly presented is better than none at all.

In addition to the basic alphabetical list, three other arrangements of place names are presented. The second displays localities by sources. In working with the astronomical literature a document may give one or both coordinates without naming the locality, or the locality may be illegible. It sometimes happens that a numerical table contains a latitude embedded as a parameter, and that analysis of the underlying function reveals the latitude. In such cases it is useful to determine as many localities as possible to which a particular latitude, or longitude, has been assigned. For this reason, the third and fourth sortings are provided by increasing longitudes and by increasing latitudes. In detail, the four sortings are:

  • Alphabetical, place names in the order of the Latin alphabet, but giving also the Arabic names, when they exist and are known; within the same place name, cross references are listed first, then coordinate pairs in numerical order of the source code numbers, hence in approximate chronological order.
  • By Sources, in numerical order of source code numbers, (hence in approximate chronological order), thence in alphabetical order of place names in the source. Modern coordinates are not displayed, nor are the Arabic form of the names.
  • By Increasing Longitudes. All of the longitudes borne by localities in the collection are listed in increasing order. Under the particular longitude, on the left hand side of the column is a list of latitudes, in increasing order. The name of the locality bearing the coordinates thus specified appears behind the latitude. This is an index only. The user may then consult the alphabetical listing to obtain the source, location within the source, and the Arabic name.
  • By Increasing Latitudes, as above except for “longitude” read “latitude” and conversely.


The project began in 1959 with support provided by the National Science Foundation, Washingtion, D.C., as an attempt to investigate relations between zījes. Some thousands of coordinate pairs were collected by Mr. Fuad I. Haddad at the American University of Beirut. At the suggestion of the late Professor William Prager, the data were transferred to punched cards and sorted and printed by the IBM-7070 computer then at Brown University, Providence, Rhode Island. At that time, such machines were not nearly as powerful and versatile as they have since become. Only capital letters, the ten decimal digits, and a few punctuation marks and special symbols could be stored and printed. The effects of these restrictions are evident in the present collection.

It soon became clear that astronomical tables were only part of a much larger corpus relevant to medieval mathematical geography, and purely geographical works were included. In 1976, the operation was transferred to the American Research Center in Egypt, supported by a grant from the Smithsonian Institution. There, and elsewhere several sources were supplied by Professor David King. At the computer center of the American University in Cairo the punched cards were replaced by magnetic tape. An additional displacement, in 1978, brought the collection back to the American University of Beirut, where Mr. Jihad Mukaddam had the data put into a machine language compatible with the new computer, and subsequently supervised programming. At this time, advice in statistical matters was obtained from Professor Mary Regier, and Dr. Reinhard Wieber supplied correlations between Greek and Arabic place names.

Since 1984, at the invitation of Professor Fuat Sezgin, whose counsel and support have been invaluable, we have worked at the Institut für Geschichte der Arabisch-Islamischen Wissenschaften at Frankfurt. From the Institute library, additional sources have been added to the collection. The results owe much to the efficient and meticulous programming of Mr. Ulrich Dorenburg.

We are grateful to the individuals and institutions named above.