Illustration Gallery

Astronomical Artifacts and Cuneiform Tablets, etc

The illustrations on this page have been compiled from a variety of sources. If advised that copyright has been infringed I will immediately remove the particular illustration(s).

Return To Section Index Page


K: Pre-Islamic Arab Constellations

21: Arabic anwā' tradition


Section of early 13th-century (Syro-Egyptian) astrolabe by Abd al-Karim al-Misri with pictorial representations of  the 28 lunar mansions and 12 signs of the zodiac. Astrolabes were important for predicting star positions. In pre-Islamic times, the lunar mansions were introduced from Hindu astronomy and amalgamated with the Anw' calendar system. Under the amalgamation, many names of prominent Anw' stars were attached to the lunar mansions.


 The pre-Islamic anwā' tradition of the Arabian Peninsula is concerned with a system of marking divisions of the year. The anwā' stars/constellations, a group of twenty-eight individual stars and small constellations along the ecliptic, constituted a calendar and meteorological system. Considerable, but fragmentary, data exists to enable reconstruction of some of the basic star calendars of the Arabian Peninsula. Before their contact with Greek-based astronomy through Arab-Islamic civilisation the pre-Islamic Arabs has their own folk astronomy. They knew the fixed stars and asterisms and used a number of fixed stars and asterisms, the so-called  anwā', for a variety of purposes. After the introduction of Islam in the 7th-century CE a substantial amount of poetry, proverbs, legends, and folk science was written down in Arabic texts. Some attention was focused on the star lore of the pre-Islamic and early Islamic Bedouins and farmers of the Arabian Peninsula. Much Bedouin star-lore has been preserved in proverbs. Specifically, from the 9th-century onwards Arabic-Islamic lexicographers and philologists collected old Arabic folk astronomy in books called Kutub al-anwā' (Books of the anwā') (Note: Most modern scholars simply write anwa'). From these books more than 300 old Arabic names for stars and asterisms (in old Arabic and early Arabic-Islamic poetry) have been recovered. Some of these star names  are likely inherited from Babylonian astral lore.

Some present-day individuals still retain knowledge of traditional Arabic star lore. Yemeni farmers till use a calendar based on the conjunction of the new (crescent) moon and the Pleiades. (This particular calendar is also appears in the recorded pre-Islamic lore.) In the Gulf region the Canopus calendar, generally based on arbitrary 10-day units from the late summer heliacal rising of Canopus, has long been a traditional calendrical system for Bedouins and sailors.

Daniel Varisco ("Stars and Texts in Arabia." (Essays from Archaeoastronomy & Ethnoastronomy News, Number 16, June Solstice, 1995)) states: "Individual stars and asterisms were used for defining the directions of the winds, timing of rain [availability of water], planting crops, pastoral activities [extent of pasture], pearling, and fishing seasons, and the like. In addition, some savants cited the location of stars as indicators of the approximate location of Mecca ...."

The pre-Islamic groups of Bedouin Arabs (i.e., the nomadic desert dwelling tribes of the Arabian Peninsula) had their own (non-standardised) individual names for the various bright stars. Indeed a rich tradition of popular star lore and seasonal lore originated on the Arabian Peninsula long before the introduction of Islam. (Paul Kunitzsch believes that the main body (younger group) of pre-Islamic Arabic star/asterism names were probably formed in the period 500-700 CE.) Pastoral tribesmen and farmers noted the risings and settings of certain stars to mark the rain periods, important winds, and seasonal events in nature. Also, they commonly regarded single stars as representing animals or people. Generally, bright stars from 1st to 3rd magnitude have had proper names in the Arabic world. The year was defined by basic seasonal change over the course of  12 months. Telling time was seasonally adapted to local contexts. An early Islamic source for the anwa' traditions, the early Islamic scholar Ibn Qutayba, states that the Bedouin Arabs of the Arabian Peninsula divided up the year from what they knew locally about the time of weather changes (hot and cold weather) and pastoral changes (appearance and disappearance of plants and pasture).

The early Bedouin asterisms greatly affected later Islamicate star names and eventually "modern" Western star names. Early Islamic astronomy was a mixture of pre-Islamic Arab star lore, and Indian, and Persian, and later Greek astronomy.

The anwa' stars (i.e., stars which marked the beginning of periods of time called anwa') are a pre-Islamic system of season- and weather-prediction by using selected prominent stars or groupings of stars. (The most important type of literature on pre-Islamic weather and astronomy is that of the anwa' dealing with potential times of rain linked with the risings and settings of certain stars.) As well as defining the seasons they act as "agricultural markers" for timing the agricultural activities. The term anwa' has the sense of rain hence the anwa' stars were associated with the rain periods. They also used them for orientation in nightly desert travels. However, the anwa' stars of the Arabs were primarily used as markers of rain. As there were variations in rainfall patterns in different parts of the Arabian Peninsula there were variously 4 to 8 periods identified. Generally, they began the year with the autumn rain followed by a sequence of rain periods.

The pre-Islamic anwa' tradition consisted of: (1) a knowledge of the risings and settings of the stars, (2) their association in particular with the settings of asterisms and simultaneous heliacal risings of other asterisms, (3) and their identification as markers of the beginning of the anwa' periods of time. (The singular term naw' (plural term anwa') is generally defined as the dawn setting of a star or asterism in the west at the same time as an opposite star or asterism rises with the sun in the east.)

This particular tradition has ultimately influenced the naming of individual stars in Western constellations. Al-Sufi's book Kitab suwar al-kawakib, our best authority for post-Islamic Arabic star-names and constellations, also included the pre-Islamic folk tradition of Arabic star names. However, al-Sufi used anwa' texts from the Islamic period.

In the post-Islamic textual traditions two systems of the anwa' have survived. The most well-known system is that of the 28 anwa' equivalent to the lunar stations of Arab astronomy and astrology with the 28 anwa' representing stars or asterisms along the zodiacal belt (ecliptic) of the celestial sphere. The origin of the system of 27 or 28 "lunar mansions" (i.e., a lunar zodiac) was the topic of considerable debate among 19th-century Orientalists. Scholars were generally divided between those who argued for a first appearance in India and those who argued for a first appearance in China. (A system of lunar mansions was established in China, India, and Sassanian Iran prior to the establishment of Islam.) A number of early Arab texts stated that a system of "lunar mansions" had formed a major calendar of the pre-Islamic Arabian Bedouin. However, there is little doubt that the Arabs borrowed the concept of the "lunar mansions" and later Islamic scholars attached seasonal almanac lore onto this imported calendar system. Julius Wellhausen (Reste Arabischen Heidentums (1897)) correctly concluded that the concept of 28 "lunar mansions" was borrowed from the astronomy of the Hindus and merged with elements of Arab star lore after the introduction of Islam. The peoples of the pre-Islamic Arabian Peninsula only recognised the seasons and rain periods of importance to them and had no concern with arbitrarily dividing the years into periods. Exactly  when the pre-Islamic Arabs received the system of 28 "lunar mansions" (manāzil al-qamar) from India, and how this was achieved, is unknown. When it was done each "lunar mansion" was identified with one of the anwa' stars or asterisms.

It is usually stated that eventually the folk tradition of Arabic star names was preserved as the "lunar mansions." This would appear to be erroneous. The system of "lunar mansions" are a type of almanac for seasonal activities. Daniel Varisco ("Islamic Folk Astronomy." In: Selin, Helaine. (Editor). (2000). Astronomy Across Cultures. (Pages 615-650).) states: "The claim that the formal model of twenty-eight lunar mansions originated as a set sequence of asterisms from a pre-Islamic star calendar cannot be sustained." The Arab-Islamic concept of "lunar mansions" appears to have been borrowed from India. (Knowledge of the Indian lunar zodiac may have existed in the Arabian Peninsula in the late 4th- or early 5th-century prior to the birth of Muhammad. (The system of "lunar mansions" is mentioned in some old Arabian poems that are pre-Koranic; and are also mentioned twice in the Koran (dated 7th-century CE.) In the post-Muhammad period, during the Arab expansion, Arab scholars began to assimilate the sciences in Iran and India.) The "lunar mansions" are the constellations/asterisms which the moon moves through in its cycle around the earth and so the "lunar mansions" are a method for dividing up the celestial sphere. However, there would have been numerous locally relevant anwa' star lists throughout the pre-Islamic Arabian Peninsula. Knowledge of only a few stars is needed to set up a seasonal calendar. Also, the 28 stars/asterisms comprising the "lunar mansions" are not evenly spaced along the moon's path (but on average are about 13 degrees apart and related in an approximate way to the moon's nightly change of position on its eastward movement through the fixed stars). (Most of the "lunar mansions" consist of pairs of stars or small groups of closely spaced stars.)

After the spread of Greek-based astronomy in the Arab-Islamic civilisation (i.e., from the 7th-century CE onwards) Arab-Islamic astronomers became familiar with the "lunar mansion" system and gave exact identifications of the stars/asterisms of the "lunar mansions" from among the stars listed in Ptolemy's star-catalogue in the Almagest. The "lunar mansions" were not actually used by the Arab-Islamic astronomers in their astronomical work. They were mostly used in astrology. The astrological texts that contained them were later to be translated into Latin and introduced into medieval western Europe.

Folk astronomy also flourished within Islamic society. This consisted of: (1) a knowledge of the sun's movement through the 12 zodiacal signs, (2) associated meteorological and agricultural phenomena, (3) use of the phases of the moon for time-reckoning, and (4) using shadows by days and the "lunar mansions" by night for simple time-reckoning.

The pre-Islamic Arabs did not expect rain at the risings and settings of all stars excepting those called naw'. The traditional anwa' texts record a length from 1 to 7 days for each naw'. Within the later "lunar mansions" scheme the length of each naw' is 13 days. These schematic lengths can hardly refer to the time for rain as several of the stations occur when there could be no rain. The sequence of rain periods attributed to the pre-Islamic Arabs varies from 4 to 8 periods of rain. It is no longer possible to reconstruct the anwa' names for the pre-Islamic seasonal sequence of rains in the Arabian Peninsula with precision.

Appendix 1:

Some pre-Islamic Arabian constellations have the same Babylonian source as the Greek zodiacal signs, but were independently inherited from Mesopotamia well before the translation of Greek scientific texts into Arabic. By Islamic times, some of these borrowed constellations had been relocated to different parts of the sky. While the classical constellation Aquarius is in its (modern) place on the ecliptic, a pre-Islamic Arabian constellation Aquarius called 'the bucket' also appears in Islamic astronomy, just to the north in Pegasus. Other zodiac signs have remained in the same place as their Greek counterparts.

Some of the Arabian versions of zodiacal constellations were (or became) far larger than the Greek counterparts, also inherited from Mesopotamia. An example is the "huge Arabic lion," a precursor of Leo. The early Arab constellation of the lion was much larger than the Aratean Leo. It sprawls across 7 classical constellations. The hind legs are in Bootes (north of the ecliptic), Virgo and Corvus (south of the ecliptic), and the forelegs are in Gemini. The lion's nose and mouth are in Cancer, while its eyes, forehead, heart and hackles are in Leo.

 In his book Kitāb al-tafhīm (= The Elements of Instruction in the Art of Astrology) the Islamic Persian astrologer and philosopher Al-Biruni (973-1048 CE) discusses some of the early Arabic constellations. He states that one of the most important was ‘Al-Asad,’ the Arabic Lion, an enormous lion constellation extending over one-quarter of the sky. He states (Section 163): "[T]heir Lion is fashioned out of some five [7?] constellations, only the eyes, forehead, neck and shoulders and the tail-tuft belong to Leo, while they make the foreleg out of the heads of Gemini, the other [leg] out of Canis minor, the nose out of Cancer and the hind legs of the two simaks [Arcturus and Spica]. In Section 164 Al-Biruni states: "With regard to the Mansions of the Moon, the Arabic lion extends from the 7th to the 14th mansions. The 7th is the foreleg; the 8th the nose; the 9th, the eyes, the 10th, the forehead, the 11th, the mane (even though it is located in the hindquarters of Leo), the 12th, the tip of the tail, the 13th, Al-Biruni says are dogs barking after the lion, the 14th, the hind legs of the lion."

All 12 constellations that later formed part of the zodiac were established in Mesopotamia by the late 2nd-millennium BCE, and appear in the Mul.Apin series (dating to circa 1300 BCE according to the astronomer Bradley Schaefer). Some of these may have been adapted in the Arabian peninsula before the 12-constellation zodiac (of 30 degree divisions along the ecliptic) was developed in the 1st-millennium BCE. As traditional Arabian astronomy includes a large lion, it can be assumed that the transmission to Arab folk astronomy occurred sometime later than 1300 BCE. Pre-Islamic folk astronomy in the Arabian Peninsula borrowed 2nd-millennium BCE stellar groupings from Mesopotamia prior to the Mesopotamian reworking circa 5th-century BCE to create an evenly spaced 12-constellation zodiac. During the late 2nd-millennium BCE the astronomical knowledge summarised in the Mul.Apin series had spread to the Middle East, Greece, Iran and India.

Paul Kunitzsch (Kunitzsch, Paul. (1986). "Remarks on Possible Relations Between Ancient Arabia and the Neighbouring Civilizations, as Found in Some Old Star Names." In: Pre-Islamic Arabia (Studies in the History of Arabia, Proceedings of the Second International Symposium on Studies in the History of Arabia ... April, 1979, Volume II, Pages 201-205)) has explained that out of a list of just over 300 individual indigenous Arabic star names only about 15 names plus a number of constellation names belonging to the zodiac may have been borrowed from elsewhere (likely Babylonia). The system of 28 lunar mansions was borrowed from India. The zodiacal constellations came from Babylonia, before or being completed with Greek influence. The introduction of the 12 zodiacal constellations (but not the Babylonian 'system' of use) did not likely occur as a single transmission. The indigenous Arabs of pre-Islamic Arabia gave 6 of the zodiacal constellations new names and 4 of the adopted zodiacal constellations were relocated in the sky, and some were made larger. Paul Kunitzsch explains that when adopting the system of 28 lunar mansions, the indigenous Arabs transferred the names of their anwa' system of corresponding pairs of stars to the lunar mansions, according to their localization in the sky. The translations of scientific works of the Indians into Arabic reportedly occurred in the 2nd half of the 8th-century CE, under the, Abbasid caliph al-Mansur. It can be assumed that the Arabs obtained their knowledge of the lunar mansions earlier than this time.

Appendix 2:

American Journal of Philology, Volume XIII, Number 51, 1892, Pages 377-378


"The first attempt to construct out of 33 or 36 Babylonian normal stars a lunar zodiac composed of 24 ecliptic stars was due to Fritz Hommel [a German Orientalist] of Munich, in the last decade of the 19th-century. He compared the Babylonian stars with the Arabian manāzils, found agreement in respect of 16 stations and concluded the manāzils were based on the more ancient Babylonian scheme. (Hindu Astronomy by P. Sharma (2004, Page 20)." Though initially given a favourable reception (due to belief in common origin between all systems), a number of important objections were soon raised against the idea of Babylonian origin. The 3 key objections were: (1) agreement only really existed for about one-third of the total number i.e., approximately 12, (2) common conspicuous stars/asterisms near the ecliptic were naturally used for the same tasks, and (3) the Arabic, Iranian, Indian, and Chinese lunar zodiacs consisted of 27/28 stars/asterisms - not 24. (In the case of China there may have originally been 23 or 24 stars/asterisms marked out along the celestial equator.) In 1989 B. Subbarayappa wrote: "I myself have always wanted to believe that the original circle of lunar mansions round the equator was Babylonian. The only difficulty is that it may be rather hard to find anything in Babylonian astronomy which could really have given rise to the hsiu and nakshatra systems."

Appendix 3:

The astronomer Bradley Schaefer has found the best estimate for the formation of the Arabic 'lunar mansion/lodge' system is 200 BCE with a statistical uncertainty of roughly 1000 years.

Copyright 2007-2014 by Gary D. Thompson


Return to top of page.

This Web Page was last updated on: Thursday, February 13, 2014, 11:00 pm.

This Web Page was created using Arachnophilia 4.0 and FrontPage 2003.

You can reach me here by email:

Return To Site Contents Page