The Fragment of Al-Kindī's Lost Treatise on Observations of Halley's Comet in A.D. 837

See Loth O. , “Al-Kindī als Astrolog”, Morgenländische Forschungen, 1875, 261–309.

There are four main sources for al-Kindī's life and works: Ibn al-Nadīm, al-Fihrist, ed. by Tajaddud (Tehran, 1971), 315–20; Ibn Ṣā^cid al-Andalusī, al-Ta^crīf bi ṭabaqāt al-'umam, ed. by Awwal Jamshid Nezhad Gh. R. (Tehran, 1997), 218–21; Ibn Abī Uṣaiba^ca, ^cUyūn al-anbā' fī tabaqāt al-aṭibā', ed. by Reza Nazar (Beirut, [?]), 285–93; Ibn al-Qifṭī, Tārikh al-ḥukamā', ed. by Lippert J. (Leipzig, 1903), 376–8.

See Adamson P. , Great medieval thinkers: Al-Kindī (Oxford, 2007), 4.

See “al-Kindī” by Jolivet J. Rashed R. , The encyclopedia of Islam (EI²) (Leiden, 1974), v, 122–3; “al-Kindī” by Jolivet J. Rashed R. , in Dictionary of scientific biography, xv, 261–7; Gutas D. , Greek thought, Arabic culture (New York, 1999), 117–20, 145–7; Adamson, op. cit. (ref. 3), 3–20.

Al-Kindī's contributions to philosophy and optics have been studied to some extent. See Rashed R. Jolivet J. (eds and transl.), Oeuvres philosophiques et scientifiques d'al-Kindī, ii: Métaphysique et cosmologie (Leiden, 1998); Adamson, op. cit. (ref. 3). However, his treatises on astronomy and meteorology deserve to be investigated comprehensively. For a report on these treatises see Sezgin F. , Geschichte des arabischen Schrifttums, vi (Leiden, 1978), 151–5, and vii (Leiden, 1979), 241–61; Rosenfeld B. A. Ihsanoglu E. , Mathematicians, astronomers, and other scholars of Islamic civilization and their works (7^th–19^th c.) (Istanbul, 2003), 39–41.

See Ibn al-Nadīm, op. cit. (ref. 2), 319. Kitāb risālatih is a standard expression repeated by Ibn al-Nadīm for every work of al-Kindī. The titles should probably be Risāla fī mā raṣada… and Risāla fī kawkab…

For a scientific and folklore terminology of comets in Islamic astronomy see Kennedy E. S. , “Comets in Islamic astronomy and astrology”, Journal of Near Eastern studies , xvi (1957), 44–51, reprinted in Kennedy E. S. , Studies in the Islamic exact sciences (Beirut, 1983), 311–18; Cook D. , “A survey of Muslim material on comets and meteors”, Journal for the history of astronomy, xxx (1999), 131–60, p. 132.

See Rada W. S. , “A catalogue of medieval Arabic and Islamic observations of comets during the period AD 700–1600”, Zeitscrift für Geschichte der Arabisch-Islamischen Wissenschaften, 1999/2000, 71–91, p. 79.

The newly-found text could be a part of al-Kindī's second treatise rather than the first one. In any case there is no evidence to decide about this issue.

Nayāzik is the plural of Nayzak [spear] which is originally an old Persian term that was used in Arabic, mostly for meteors or comets, see Kennedy , op. cit. (ref. 7), 311. For its use as a word for comet see also Ibn Hibintā, al-Mughnī fī aḥkām al-nujūm, facsimile edn (Frankfurt, 1987), ii, 134. In some later Persian treatises on meteorology, it was used to express a kind of opti-atmospheric phenomenon as well. However, comets were usually called as “a star with tail/tails [kawkab dhawāt al-adhnāb]” or “a star with locks of hair [kawkab dhudhu'āba]” in Arabic literature. See also Cook, op. cit. (ref. 7), 132.

This treatise appears in a collection of manuscripts published online by Princeton University. See http://pudl.princeton.edu/objects/r781wg07m (treatise no. 47).

MS 373y, fols. 237r-248v, Princeton University Library. See also Sezgin, op. cit. (ref. 5), vii, 372.

Ḥunayn led a famous circle of translators in Baghdad at the time of Mutiwakkil (reigned 847–61). Ḥunayn himself translated numerous Greek works in the various scientific fields into Arabic; however, the majority of them were in medicine. See Sezgin F. , Geschichte des arabischen Schrifttums, iii (Leiden, 1970), 247–56.

See also Bos G. Burnett C. , Scientific weather forecasting in the Middle Ages: The writings of al- Kindī. Studies, editions and translations of the Arabic, Hebrew and Latin texts (London and New York, 2000), 201, 359–60.

This passage is in fols. 2 40r-41v.

Al-Maqdisī, MS 178 (mīqāt), fols. 6r-6v, Library of Dār al-Kutub al-MiṢrīyya. For a report on the surviving manuscripts of this treatise, see King D. , A survey of the scientific manuscripts in the Egyptian National Library (Cairo, 1986), 105, 212, 304.

Various appearances of Halley's Comet were recorded in the Islamic sources, see Cook, op. cit. (ref. 7), 131–60.

Ibn al-Athīr, al-Kāmil fial-tārīkh (Beirut, 1398/1978), v, 245, reports that a comet appeared in 222 A.H/A.D 837 and it became visible to the naked eye for about forty nights (for analysis see refs 38 and 44). See also Rada, op. cit. (ref. 8), 78 and Cook, op. cit. (ref. 7), 137–8. For the reports on the 837 return of Halley's Comet from other nations, see Kronk G. , Cometography, i: Ancient-1799. A catalog of comets (Cambridge, 1999), 125–7.

Seargent D. , The greatest comets in history (New York, 2009), 44.

This sentence could be translated as: “Before that, it travelled less than one degree on one night and more than one degree on another”, based purely on the text. However, the computations revealed what al-Kindī means.

The text incorrectly gives this value equal to 12°.

The text gives the incorrect value as 30°.

The text incorrectly gives this value equal to 12°.

This Program has been specialized to compute the various coordinates of the comets and minor planets and it is accessible through the following database: http://ssd.jpl.nasa.gov/?horizons. Here I have rounded the computed values to 0.01°. The accuracy of the original values given by the computer program is equal to 0.0000001°. Obviously such a precision is far beyond of values that could be obtained by the ancient astronomers. On the other hand, with regard to the modern computations, the passage of Halley's Comet from its perihelion has not yet been determined more precisely than 0.05^d ≈ 1.2^h for A.D. 837. Thus such a difference can produce errors in the comet's positions equal to 2° when the comet was near the Earth. Thus the results of the computer program are not practically meaningful to such degree of precision.

See Kronk , op. cit. (ref. 18), 125–7.

See Kiang T. , “The past orbit of Halley's Comet”, Memoirs of the Royal Astronomical Society , lxxvi (1972), 27–66; and Kronk, op. cit. (ref. 18), 125–7.

The Chinese astronomers located the comet's positions in their lunar mansions system (for Chinese lunar mansions and its origin see Chu C. , “The origin of twenty eight mansions in astronomy”, Popular astronomy , lv (1947), 62–77; Yampolsky P. , “The origin of the twenty-eight lunar mansions”, Osiris, ix (1950), 62–83). For a list of the Chinese observational records of the appearance of Halley's Comet in 837, see Kiang, op. cit. (ref. 26), 52, and Kronk, op. cit. (ref. 18), 125–7. Kiang (table XI, col. 5) tried to identify the comet's right ascensions based on the given positions by Chinese astronomers, but it is unclear how Kiang could reproduce the right ascensions in precision equal to 0.1° based on the Chinese imprecise records. It is obvious that in the Chinese celestial coordinate system, the angular distance of the comet is given in relation to a single star (or a group of stars), each of which indicated a lunar mansion. On the other hand, Kiang has not presented even approximate values of the comet's declinations based on the Chinese texts rather has independently calculated the declinations based on the computed orbital model. However, I have adopted those given equatorial coordinates (observed right ascensions and those computed declinations) by Kiang and converted them into the ecliptic coordinates in order to enable us to compare these Chinese coordinates with al-Kindī's in the ecliptic coordinates system (see ephemeris table).

In this analysis several parameters such as the comet's rising time, sunrise-sunset times and the comet's altitude have been taken into account. In the ephemeris table the times are given in UT, thus for converting them to the local time (LT) in Baghdad, +3 should be added to the given hours.

See Seargent, op. cit. (ref. 19), 44–5.

See Kronk, op. cit. (ref. 18), 125.

It is also possible that al-Kindī had read for Ḥunayn the comet's parameters based on his observational diaries. Some obvious numerical errors in the text indicate the possibility of such hypothesis; but it may also imply the long interval between the records' appearing in the existing recension of Nayāzik compared to that of al-Kindī's original treatise, and it also seems that the original values were in the abjad system rather than the Indian numbers. However, it is hard to believe that al-Kindī cited the mentioned parameters by memory during a simple conversation with Ḥunayn.

Al-Maqdisī, op. cit. (ref. 16).

It should be kept in mind that the dates are not in the text but result from my computations.

It seems that al-Kindī measured the daily motion of the comet as being less than 1° over this period according to the text.

See Yeomans Kiang , op. cit. (ref. 20), 636.

The Arabic manuscript probably originally had numbers in the abjad system, although they now appear as Indian numbers; for instance, the value of 52 (nab) was incorrectly read as 12 (yab) because of the similarity in writing the two abjad numbers. As for the Chinese records, it can be said that the records from 5.8 to 8.8 April show great discrepancies compared to those derived from HJPL, as well as those values given by al-Kindī. This may originate either from inherent imprecision of Chinese records (due either to miscopying or to limitations in the original observations) or the method of coordinate conversion made by Kiang, op. cit. (ref. 26), 52. My conjecture is that it mainly originates from the technique by which the comet's positions (in the lunar mansions system) were converted into right ascensions, because it is unlikely that Chinese astronomers recorded the comet's positions so imprecisely. Moreover, the Chinese records on 4.8, 10.7 and 14.6 April are relatively precise observations compared to the results of the computer program. As a whole, it can be said that Al-Kindī's ecliptic coordinates show better accordance with the computed values derived from HJPL than the coordinates interpreted and given by Kiang for the Chinese observations.

Since Ibn al-Nadīm, op. cit. (ref. 2), 316, attributes to al-Kindī a treatise on construction and using of armillary sphere, it is possible that al-Kindī made his observations with an armillary sphere. We have no information on the accuracy of al-Kindī's original observations. Only in one case, al-Kindī points out that the ecliptic longitude of the comet was greater than 20° of Capricorn by several minutes of arc; however there is no mention of the precise value.

The comet's distance from Earth was 0.68 AU at this time. See Kronk, op. cit. (ref. 18), 125. According to Ibn Athīr's reports, op. cit. (ref. 18), the comet was visible for forty nights, so if we assume that the Muslims spotted the comet at the same date that Chinese did (22 March), it would be last seen around 1 May and that does agree with the Chinese report.

According to Ibn Hibintā's citation, op. cit. (ref. 10), 140, such criterion originated from Kanka (the astronomer who came from India and served at the court of the Calif Hārūn al-Rashīd who ruled from 786 to 809). Moreover, Ibn Hibintā, op. cit. (ref. 10), 134, quotes an anonymous commentator as saying that the comets appear when the Mercury's elongation from the Sun is equal to 17°. This shows that some astronomers or astrologers sought criteria for determining apparition of comets in the early period of Islamic astronomy.

Indeed, it is unclear from whom is derived the observation of Halley's Comet in 912 based on the treatise of Nayāzik. However, al-Maqdisī, op. cit. (ref. 16), in his Tuḥfat al-albāb, asserts that this report is from al-Mughnī's author (īb al-Mughnī). We know that Ibn Hibintī (d. after 317/929) is the author of the al-Mughnī fī aḥkām al-nujūm, although there is no trace of this report in the surviving incomplete text of the al-Mughnī.

For an analysis of variations of the orbital elements of the Halley's Comet over history, see Yeomans and Kiang, op. cit. (ref. 20), 643.

See Kronk, op. cit. (ref. 18), 125.

Ibn al-Athīr, op. cit. (ref. 18). Obviously Ibn al-Athīr himself (d. 1233) was not an eyewitness of the comet's passage in 837 and the sources from which he has quoted this report are to us unclear. Thus it is impossible to judge about the geographical location where the comet first appeared left of its qibla direction. However, the comet could be located in a southwest direction of the local horizon in different parts of the Islamic region such as Iraq, Iran, and eastern parts of Saudi Arabia. Moreover, it should be noted that there is a contradiction in Ibn al-Athīr's report, because he adds that the comet lasted for forty nights. If the comet was spotted by the naked eye in the western horizon before attaining its perihelion sometime in the first week of February, forty nights later it must have been seen for the last time around 20 March, which is not correct. But by adding forty nights to 22 March, one reaches the end of April which agrees with the last date (i.e. on 28 April) that the comet was seen by Chinese astronomers (see ref. 38).

On 20 Jan. the comet's distance from Earth was about 1.5 AU.

For this formula see Hughes D. W. , “Temporal variations of the absolute magnitude of Halley's comet”, Monthly notices of the Royal Astronomical Society, cciv (1985), 1291–5.

It is important to remember that comets normally appear brighter after passing perihelion compared to the same distance before perihelion.

Indeed HJPL estimates an even greater value for the comet's total magnitude (i.e., the comet's coma was fainter), as being equal to 5.8 at this time.

It should be noted that the comet's passage of 837 was in such a way that the comet reached higher altitudes in the morning sky after passing its perihelion.

There is other hypothetical or observational evidence on outburst phenomenon in various past returns of the Halley's Comet. See Seargent, op. cit. (ref. 19), 47, 49, 54–5, 58–9, 63. For research on the possible causes of Halley's outbursts at large heliocentric distances, see P. Gronkowski, “Outbursts of comets — The case of 1P/Halley”, Planetary and space science, 1 (2002), 247–56.

See Kronk, op. cit. (ref. 18), 125.

This computation is correct if one assumes that the comet's brightness was constant between 22 March and 10 April. However, the total magnitude of −3.5 is in agreement with Seargent's statement, op. cit. (ref. 19), 45, that the comet's maximum total magnitude for the passage of 837 must have been between that of Jupiter and Venus, but most probably closer to the latter. However the HJPL estimates that the maximum total magnitude was equal to −1.5.

See Seargent, op. cit. (ref. 19), 4.

See Cook, op. cit. (ref. 7), 135–7.

See Kronk, op. cit. (ref. 18), 146–50.

The following is Ibn al-Jawzī's report (see Ibn al-Jawzī, al-Muntaām fī tārikh al-'umam wa'l-mulūk (Beirut, 1412/1992), xiii, 123) on the apparition of three comets (?) in 912: “Three stars with tails appeared. One of them appeared on the night of 25 Ramaḍān [14 May 912] in Leo. The second of them appeared on Tuesday, 11 Dhū al-Qa'da [30 June] in the east, and the third appeared on the night of Wednesday, 20 Dhū al-Qa'da [8 July] in Scorpio. It [?] stayed a number of days and then gradually died out” (cf. Cook's translation and calendar conversion, op. cit. (ref. 7), 139). It seems that Ibn al-Jawzī's report has not been properly examined. His first record on the apparition of a comet on 14 May is interestingly in accordance with the Chinese report giving the position of a comet in Leo on 15 May. Apart from the large distance of Halley's Comet from both Sun and Earth on these dates, modern computations show that its ecliptic longitude was 42°. Thus Halley's Comet was not in Leo as constellation or as a sign at the time. As for the second report, it is in agreement with our newly found record and it definitely concerns the apparition of Halley's Comet (see above). However, the third report has nothing to do with Halley's Comet's coordinates and trajectory: its ecliptic longitude on 10 July 912 was about 85° rather than being between 210° and 240°, corresponding to the sign of Scorpion. (Cook, op. cit. (ref. 7), 139, links Ibn al-Jawzī's third report to an appearance of Halley's Comet.).

See ref. 58.

The value of H₀ was not computed in Hughes's research for the passage of A.D. 912. See Hughes, op. cit. (ref. 46), 1292, Table 1. This was probably due to controversial reports of the Halley's Comet for this passage. However, here I have computed the comet's apparent magnitude based on the mean values of H₀ and n given by Hughes (p. 1291) as 5.49 and 5.15 respectively.

See Kronk, op. cit. (ref. 18), 146–50.

See T. Heidarzadeh, A history of physical theories of comets: From Aristotle to Whipple (New York, 2008), 34. As for the Islamic astronomy, another report known from its late period deals with the comet of 1825. It was observed by Ghulām ūssain Jawnpūrī in India, who gives several ecliptic coordinates of this comet (see S. M. R. Ansari, “Practical astronomy in Indo-Persian sources”, Indian journal of history of science, xxxvii (2002), 255–66, pp. 257–8). It is unclear whether or not Jawnpūrī was inspired by the modern astronomical sources in doing such measurements, although there are some traces of modern European astronomy in his works.

See P. Lettinck, Aristotle's meteorology and its reception in the Arab world (Leiden, 1999), 74–96.

See Rada, op. cit. (ref. 8), 71–91, and Cook, op. cit. (ref. 7), 131–60.

See Pliny [the Elder], Natural history, ed. and transl. by Rackham H. (Cambridge, MA, 1971), i, Book 2, 235.

Aristotle criticized theories that dealt with comets as heavenly bodies in his Meteorology (see Aristotle, Meteorologica, transl. by H. D. P. Lee (Cambridge and London, 1952), 38–48). We now know that some later astronomers and philosophers from the Islamic period developed theories considering the comets as celestial objects (or supposed them under the effect of the celestial forces) rather than being merely among the phenomena appearing in the Earth's atmosphere (see Abu'l-Barakṭ al-Baghdādī, al-Mu ^ctabar fi'l-hikma (Haydar-Abad, 1358/1939), ii, 233; Ibn Hibintī, op. cit. (ref. 10), ii, 134; L. Thorndike, “Albumasar in Sadan”, Isis, xlv (1954), 22–32, p. 29). However, such criticisms did not lead to the abandonment of Aristotle's cometary theory in the medieval period.

For the contribution of Western astronomers to cometary theories in early modern science, see Heidarzadeh, op. cit. (ref. 75), 36–45.