Lot 123
  • 123


200,000 - 300,000 GBP
286,000 GBP
bidding is closed


  • 18.4cm. diam.
cast brass, with delicate silver and gold-inlaid foliate decoration to front of cusped throne, the reverse of the throne with corresponding decoration but with the inlay lost, the entire circular edge with a scrolling vine motif in silver, containing four plates (one of which is original), later, probably Ottoman rete, alidade, and suspension loops


Private Collection, UK, since 1977.

Catalogue Note


signed by Muḥammad ibn Abi Bakr al-Qawwas, made in Syria in 752 AH [1351/52 AD]

Original remaining sections include the mater and one plate

The instrument dates from fourteenth-century Syria, a milieu that was arguably the leading centre of astronomy in the world. The activities of the leading Syrian astronomers and instrument-makers are reasonably well documented, and it comes as something of a surprise to the few modern specialists that a fine ‘new’ astrolabe should come to light signed by a maker who is unknown to the literature. The leading astronomers of that period were Ibn al-Shatir, best known for his ‘new’ geometric models for the sun, moon and planets (which happen to be the same as the models used by Copernicus some 150 years later); al-Khalili, best known for his remarkable tables for solving all of the standard problems of spherical astronomy for any latitude (without parallel in pre-modern European astronomy); al-Mizzi, whose instruments fetched high prices already in his own time; and Ibn al-Sarraj of Aleppo, who constructed the most sophisticated astrolabe ever made (which functions for all latitudes in five different ways).

The maker of this ‘new’ piece is Muhammad ibn Abi Bakr al-Qawwas. The nisba al-Qawwas means ‘archer’, from Arabic qaws, ‘bow’, and the nisba may be inherited. Archery was a fine art in Mamluk Syria, not least because of the need to keep any Mongol or European invaders at bay. Several manuals were produced identifying the different types of bows available and their use, and some of these have been published, notably that of the Mamluk prince Taybugha al-Ashrafi al-Baklamishi al-Yunani (Laatham & Paterson, Saracen archery. See also the article ‘qaws [= bow]’ in Encyclopaedia of Islam, 2nd edn.). This notwithstanding, our Muhammad ibn Abi Bakr al-Qawwas is unknown to the medieval and modern literature on Muslim archery (Professor Rex Smith, expert on medieval Arabic and especially Muslim archery, has kindly confirmed that this al-Qawwas is unknown to him)  and apparently also to the medieval Arabic biographical literature (Prof. Daniel M. Varisco, expert on medieval Islamic folk astronomy and agricultural practices, has pointed out that the name Muḥammad ibn Abi Bakr al-Qawwas is attested in the ‘Ilam al-nubala’ bi-ta’rikh halab al-shahba, a biographical dictionary of scholars in Aleppo by M. Raghib ibn Mahmud al-Tabbakh al-Ḥalabi (V, pp.433-4), although he is recorded as having died in 934 AH/1527/28 AD.

In addition he is known neither to the literature on astronomy and mathematics in Mamluk Syria nor to the modern documentation on Syrian instrumentation (see King, ‘Astronomy of the Mamluks’; Charette, Astronomical instrumentation in 14th-century Egypt and Syria, and Brentjes, ‘al-Sakhawi on muwaqqits, mu’adhdhins, and the teachers of astronomy in Mamluk cities’; also King, ‘Astronomical instruments from medieval Syria’). On the other hand, there are two important astronomers with deceptively similar names known from the thirteenth century (namely Muhammad ibn Abi Bakr al-Farisi, a man of Iranian extraction who worked in Aden, and Muhammad ibn Abi Bakr al-Kawwashi from Quṣ in the Nile Valley – see King, Mathematical astronomy in medieval Yemen, pp.23-26 and 27). The engraving
The maker of this astrolabe used a distinctive and singularly elegant ornamental Kufic script.

The numeral forms
In Islamic astronomy numbers were expressed in the sexagesimal (base 60) alphanumerical system known as abjad (a=1, b=2, j=3, d=4, etc., including letters for the tens and hundreds). Thus, for example, l-j k-z, stands for 33°27´. On this astrolabe, exceptionally, the date of construction is given in Hindu-Arabic numerals, which were generally used outside the realm of astronomy (Irani, Arabic numeral forms).

The throne
The front of the throne, whose outline is multi-lobed, is elegantly inlaid with silver and gold stylised floral scrolls. This is an unusual feature on Islamic astrolabes, even on Syrian ones; only the spectacular astrolabes of the thirteenth-century artisans ʿAbd al-Karim al-Miṣri (Museum of History of Science, Oxford, and British Museum, London, – see Gunther, Astrolabists, nos.103 and 104) and al-Sahl al-Nisaburi (Germanisches Nationalmuseum, Nuremberg – see King, Synchrony, XIVb: 677-684) are decorated in this way.

The mater
The mater bears no markings. The outer scale is divided into 360° into 1° intervals subdivided and labelled for each 5°. The labels run (clockwise from the top):

5 - 10 - 5 - 20 - 5 - ...  - 80 - 5 - 90 - 5 - 100 -
5 - 10 - 5 - ... - 90 - 5 - 200 - 5 - 10 - 5 - ... - 90 - 5 - 300 -
5 - 10 - 5 - ... - 5- 60.

The peg to hold the plates is offset from the vertical diameter at some 7° to the right. The cut-outs on the plates are situated accordingly.

The back
On the rims of the upper left and right quadrants of the back are two altitude scales, with labelled divisions for each 5° subdivided into divisions for each degree. Within these is a full semicircle of a sexagesimal trigonometric grid, resembling modern graph-paper. Normally a quarter-circle of such markings is found on some astrolabes. Here the maker has taken the trouble to mark the upper half of the instrument with a set of horizontal lines for each unit of the 60-unit vertical radius. These are intersected by a set of vertical lines for each unit of the double 60-unit horizontal diameter. The accuracy with which this grid has been achieved is impressive indeed. It compares with that of the grid on the back of the astrolabic quadrant for Damascus of al-Mizzi in the British Museum (On the other hand, such markings are attested already on the astrolabic equatorium of the Baghdad astronomer Hibat Allah al-Asṭurlabi dated 1120-21 AD (preserved in the Museum für Islamische Kunst, Berlin), invented by the Baghdad astronomer Abu Jaʿfar al-Khazin c.950 (King, Synchrony, vol.1, pp.71-73). The markings here are labelled below the horizontal diameter, al-jayb [al-]sittini, ‘the sexagesimal sine’ or ‘the sine to base 60’. (In medieval Islam the sine and cosine function were to base 60 rather than 1 as we use today). On the lower left rim is a scale labelled ‘asr afaqi, ‘universal (scale for the altitude of the sun at the beginning of the) ʿasr (or mid-afternoon prayer)’. One sets the alidade at the midday solar altitude on the altitude scale on the upper left and the lower end will indicate the altitude of the sun at the beginning of the prayer time, and this works for any latitude. (If H represents the midday solar altitude and A represents the solar altitude at the time for the prayer, then cot A = cot H + 1.) As we shall see, 7 and 12 were also used as bases for these shadow functions. On the lower left rim is a scale labelled zill asabiʿ, ‘shadow in digits’, indicating the shadow of cast by a gnomon length 12, that is 12 x cot h. One sets the alidade at the instantaneous solar altitude on the upper left altitude scale and the lower end will indicate the shadow. The two squares below the centre of the back are for finding the horizontal or vertical shadows to base 12 on the left and 7 on the left. They are labelled murabbaʿ al-ẓill aqdam / aṣabiʿ, ‘shadow square for feet / digits’. The horizontal and vertical scales are marked, mabsuṭ, ‘horizontal’,  and mankus, ‘vertical’.  (It should be borne in mind that all six of the basic trigonometric functions were used from the ninth century onwards in Islamic astronomy, if not in mathematics.)

The signature is in an elegant cartouche below the horizontal diameter and reads:

 ‘Constructed by Muḥammad ibn Abi Bakr al-Qawwas in the year 752 (AH = 1351-52 AD).’

No other cartouches of this kind are known from other Syrian instruments - the closest examples on an astrolabe are from fourteenth-century Toledo (King, Synchrony, XV: 841, 847, 890-891).

The one original plate
On all plates the vertical and horizontal diameters represent the meridian and the east-west line. The base circles are marked for the Tropic of Capricorn on the outside and the Tropic of Cancer on the inside, the celestial equator being between these. Each set of markings includes circles for altitude from the horizon up to the zenith and for azimuth measured round the horizon.

One plate (no.1) is original, presenting two sets of astrolabic markings. Plate 1a has a set of markings stated to be for ʿard lj l’, ‘latitude 33°30´. This is undoubtedly intended for Damascus. Both Ibn al-Shatir and al-Khalili used this value; al-Mizzi preferred 33°27´ (the accurate value is 33°•••´). There are labelled altitude circles for each 2° from the horizon up to the zenith, and unlabelled azimuth circles for each 10° around the horizon.

Plate 1b bears a set of similar astrolabic markings for latitude l-z, ‘37°’, and here it is not immediately obvious which locality might have been intended. Al-Khalili’s corpus of tables for astronomical timekeeping and regulating the prayer times contains a geographical table giving longitudes and latitudes mainly in Syria and Palestine, but includes no significant localities with latitude c.37° (King, Synchrony, II: 390-3). Even the extensive lists of geographical coordinates from medieval Islamic sources do not indicate a possible candidate (Kennedy & Kennedy, Geographical coordinates from Islamic sources). It seems that three original plates are missing and have been replaced.

Replacement parts
The rete and other three plates, as well as the alidade, are additions; the maker has made an attempt to imitate the distinctive engraving of the first. A distinct difference from al-Qawwas’ astrolabe is the late form for ‘20’ in numbers like 20+4 = 24, written as الد . This curious form of ‘k’ for ‘20’ is typical of certain Ottoman hands. However, it is by no means clear when these parts might have been added. The engraving cannot be modern, to reproduce the distinctive script of the original, neither can the astronomical markings on the three plates be modern, for even though they are not competent – the horizon does not properly intersect the horizontal east-west diameter at the east and west points – the constructor at least made them for appropriate latitudes. Since all of the replacement parts are not modern and they are clearly based on or copied from components of the original astrolabe of al-Qawwas, they merit our detailed attention.

Although the rete features star-pointers that are more or less in the right places, and most of the star-names are correct, it appears to have been copied from an original rete that was broken. On the ecliptic ring the signs of the zodiac are 180° out of place, that is, they begin on the right rather than on the left, and there is a meaningless Arabic inscription around the outside of the ring, on which there are no divisions for a scale. On the ecliptic ring the names of the signs are correctly written but there are no divisions on the outer rim. The ‘text’ on the outer part of the ecliptic ring reads:

ح ط يا يب كا كد كز ل يه ح ط يب  كد كز كا يه ح  ي يا و ح يب كد كز ل يه ح و يا يب كا ح ي يه كد كز ط ح و ل كا يه

8 9 11 12 21 24 27 30 15 8 9 12 24 27 21 15 8 10 11 6 8 12 21 24 27 30 15 8 6 11 12 21 8 10 15 24 27 9 8 6 30 21 15

The mainly meaningless numbers surely represent a very unsuccessful attempt to reproduce the sequence 6 - 12 - 18 - 24 - 30 in each zodiacal sign. Furthermore, the ’15’ is used where one would expect a border between the zodiacal signs, inevitably not properly divided into 12. The star-names on the rete are the following, most of them designating well-known astrolabe-stars. Diacritical points have been added here where necessary, giving the maker the benefit of the doubt. Names that are incorrectly written are underlined. Correct or more complete versions are given for names which are in error or abbreviated. The K-numbers are those in Paul Kunitzsch’s list of 60 astrolabe stars (Kunitzsch, Arabische Sternnamen in Europa, pp.59-96).

There are three replacement plates with six sets of standard astrolabic markings of different kinds. On three we find the meaningless inscription ʿarḍ / saʿat , ‘latitude/ hours’, with no associated numbers, whereby it should be borne in mind that most astrolabe plates show the latitude and the associated length of maximum daylight in hours and minutes. The altitude circles are for each 3° (not 2° as on the original plates) and the azimuth circles for each 10°. Two sets of markings (plates 2a and 2b) are for latitude c.22°30´, that is, Mecca, though this is not stated (On early values of the latitude of Mecca see King, ‘Earliest Muslim geodetic measurements”, pp. 225-6). The appearance of the ‘identical’ markings for the same latitude is not identical. Another set of markings (plate 3b) is for latitude c.43°. These derived latitudes have been determined from the solar midday altitude at the equinoxes (= 90° minus latitude) and, given the inexact nature of the markings, can only be considered approximate.

A plate of full horizons has been less successfully achieved. The lower half displays horizons labelled for each 3° of latitude from 3° to 90° inside the equatorial circle, and the lower half the same horizons outside the equatorial circle. The horizons are so cluttered near the two circles on the horizontal diameter representing the east and west points that they are not continuous. For lower latitudes the horizons do not even pass through the east and west points.

A plate of half horizons, which can in theory be used for operations involving only the horizon (so one does not need any altitude circles), shows four sets of half horizons serving the following latitudes, constituting a complete set from latitude 10° (to the south of Yemen) up to the Arctic Circle:

10 - 18 - 26 - 34 - 42 - 50 - 58 - 66
12 - 20 - 28 - 36 - 44 - 52 - 64
14 - 22 - 30 - 38 - 46 - 54 - 62
16 - 24 - 32 - 40 - 48 - 56 – 60

At each of the quadrants there is a scale with no divisions. It is labelled [al-]mayl al-kulli, ‘maximum declination’, that is, the obliquity of the ecliptic, c.23°30´. The arguments for the declination are presented even though there are no divisions to the scale:4, - 8 - 12 - 16 - 20 - 24. The curious ‘20’ is used on this plate so that even though the astronomical markings seems to be in order, it belongs to the spurious additional parts. There is an astrological plate labelled at the centre ‘latitude 32°’, most probably originally intended for Jerusalem, also a centre of limited activity in the fourteenth century and thereafter. The markings consist of a principal set of arcs emanating from the centre labelled with the (masculine) ordinal numbers in words for each of the 12 astrological houses thus: ṭaliʿ (ascendant, marking the beginning of the first house) - thani (second) - thalith (third) - thani ʿashar (twelfth). The intervening spaces are divided by similar arcs for each 6° of each house, labelled 6 - 12 - 18 - 24 - 60. The ’20’ of ’24’ is written as two vertical lines. The markings are incorrectly drawn and have gone quite awry at the top (See Gunther, Astrolabes, I, between pp.250 and 251, for such a plate, correctly constructed, from eleventh-century al-Andalus).

The alidade has the names of twelve signs of the zodiac written along its length, six on each sides; further proving that the maker of the additional or replacement parts was not familiar with the scientific functions of an astrolabe.

On Islamic astrolabes, see Gunther, Astrolabes of the World (1932). On astronomy in medieval Syria and numerous surviving instruments, see King, ‘L’astronomie en Syrie à l’époque islamique’, partly reworked in idem, In Synchrony with the Heavens, XIVb: ‘Astronomical instruments from medieval Syria’. On the astrolabe see most recently idem, ‘The astrolabe: what it is and what it is not’. On the latitudes associated with astrolabes see idem, Synchrony, XVI: ‘The geographical data on early Islamic astronomical instruments’.

Sotheby’s is grateful to Professor David A. King for his assistance in describing this astrolabe.


Sonja Brentjes, ‘Shams al-Din al-Sakhawi on muwaqqits, mu’adhdhins, and the teachers of various astronomical disciplines in Mamluk cities in the 15th century”, in Emilia Calvo, Mercè Comes, Roser Puig and Mònica Rius, eds., A Shared Legacy: Islamic Science East and West: Homage to professor J.M. Millás Vallicrosa, Barcelona, 2008, pp.129-150.

François Charette, Mathematical instrumentation in 14th-century Egypt and Syria – The illustrated treatise of Najm al-Din al-Misri, Leiden, 2003 (analysis of over 100 instrument-types, mainly astrolabes, quadrants and sundials).

Robert T. Gunther, The Astrolabes of the World, 2 vols., Oxford, 1932, repr. in 1 vol., London, The Holland Press, 1976.

Thomas Hockey et al., eds., The Biographical Encyclopedia of Astronomers, New York: Springer, 2007, available at http://islamsci.mcgill.ca/RASI/BEA/: standard reference on significant Muslim astronomers, containing articles on Ibn al-Shatir, al-Khalili, and al-Mizzi.

Rida A.K. Irani, ‘Arabic numeral forms’, Centaurus 4 (1955), pp.1-12, repr. in Kennedy et al., Studies, pp.710-721.

Kennedy et al., Studies in the Islamic Exact Sciences, D. A. King and M. H. Kennedy, eds., Beirut, 1983.

– & Mary Helen Kennedy, Geographical coordinates of localities from Islamic sources, Frankfurt am Main: Institut für Geschichte der Arabisch-Islamischen Wissenschaften, 1987.

David A. King
– , ‘The astronomy of the Mamluks’, ISIS 74 (1983), pp.531-555, repr. in idem, Islamic Mathematical Astronomy, London, Variorum, 1986, III.
– , Mathematical Astronomy in Medieval YemenA Bio-Bibliographical Survey, (Publica­tions of the American Research Center in Egypt), Malibu CA, Undena, 1983.
– , ‘L’astronomie en Syrie à l’époque islamique’, in Sophie Cluzan & Eric Delpont & Jeanne Mouliérac, eds., Syrie, Mémoire et civilisation, Paris, 1993, pp.386-395, and [‘Instruments astronomiques syriens’], pp.432-443 & 480, and pp.485-7; original English text expanded in In Synchrony with the Heavens, XIVb: ‘Astronomical instruments from medieval Syria’ (pp.659-724).
– , ‘The earliest Muslim geodetic measurements’, Suhayl International Journal for the History of the Exact and Natural Sciences in Islamic Civilisation (Barcelona) 1 (2000), pp. 207-241, repr. in idem, Islamic Astronomy and Geography, Aldershot & Burlington VT, Ashgate-Variorum, 2012, X.
– , In Synchrony with the Heavens – Studies in astronomical timekeeping and instrumentation in Islamic civilization, 2 vols., 1: The Call of the Muezzin, & 2: Instruments of Mass Calculation, Leiden & Boston, 2005 (vol.1 deals with astronomical timekeeping, vol.2 with instrumentation). Contains: XIIIa: ‘The neglected astrolabe – A supplement to the standard literature on the favourite astronomical instrument of the Middle Ages’; XIVb: ‘Some astronomical instruments from medieval Syria’; XVI: ‘The geographical data on early Islamic astronomical instruments’; XVIII: ‘A checklist of Islamic astronomical instruments to ca.1500, ordered chronologically by region’.
– , ‘The astrolabe: what it is and what it is not’ (2018), available at www.davidaking.academia.edu.

Paul Kunitzsch, Arabische Sternnamen in Europa, Wiesbaden, Otto Harassowitz, l959

– , ‘Al-Sufi and the astrolabe stars’, Zeitschrift für Geschichte der arabisch-islamischen Wissenschaften 6 (1990), pp.151-166, repr. in idem, Stars and Numbers – Astronomy and Mathematics in the Medieval Arab and Western Worlds, Aldershot & Burlington VT, Ashgate-Variorum, 2004, XIII.

John-Derek Latham and W.F. Paterson, Saracen archery: An English version and exposition of a Mameluke work on archery (ca. A.D. 1368), London, The Holland Press, 1970.

Oxford, Museum of the History of Science: www.mhs.ox.ac.uk/astrolabe/, webpage of the world’s largest collection of astrolabes.