and indirect in ETCSL 2. 1. 7 (Gudea Cylinders A and B) 134 f., and in ETCSL 4. 16. 1 (Hymn
Nisaba A), and is perhaps very ancient. Enlil is also a star according to ETCSL 2. 4. 2. 07 Sˇulgi
G): 1 – 8.
6 BM 22696 : 6 , cited in Rochberg 2005 : 68. ETCSL 5. 3. 6 (The debate between Silver and Copper)
refers to tablets of the stars.
7 As made clear by the many references within the omen corpus to kings of the Old Akkadian
and Ur III periods. Bibliography in Brown 2000 : 246 § 2.
8 In a catalogue of incipits from Ur, dating to the Old Babylonian period, the title Enu ̄ ma Anu
Ellilappears in both Akkadian and Sumerian (Brown 2000 : 248 § 7 ). Cf. now Rochberg 2005 :
69 on the small series of eclipse omens from Old Babylonian times.
9 Approximately half of Enu ̄ ma Anu Ellilhas been edited to date (Brown 2006 , n. 23 ).
10 Al-Rawi and George 1991 – 2.
11 The reader may justifiably ask what is meant by ‘remotely’ here. As he or she reads on, it will
become clear that the accuracy of the ideal schemes of Enu ̄ ma Anu Ellil 14 and the like were
such as to ensure that the actual date, or time, of the phenomenon could still be understood
to have been the consequence of the arbitrary will of the god in question. Later, the higher
levels of accuracy achieved would of necessity have made that understanding questionable.
12 Hunger and Pingree 1988.
13 See most recently Rochberg 2005 : Ch. 2.
14 Brown 2000 : 189 – 207 summarises the relevant data. The evidence that these scholars were
capable of producing some accurate observations and predictions, and were aware of some
characteristic periods for the planets is unequivocal. The limits of their abilities can also be
determined from the Letters and Reports, showing that this science was still in its infancy in
the seventh century BC.
15 The clearest evidence is in text DT 72 +, last treated in Brown ( 2000 : 193 ). There is also a
hint that a particular interval of 6 , 585 days and ca. 8. 5 hours, used to determine not only the
date of an eclipse but its time, the so-called Saros, was in use in the seventh century BCand
perhaps even the mid-eighth (Brown 2000 : 205 ; Steele 2002 ).
16 The surviving ‘Astronomical Diaries’ come almost entirely from Babylon and represent a small
part of a database that in its complete form dated from the mid-eighth century BCuntil
perhaps as late as the first century AD. Most are published in Sachs and Hunger 1988 – 96.
The Diaries record not only observations of astral phenomena such as planetary phases, but
weather phenomena, the level of the Euphrates, and some historical data. Many of the
astronomical data were calculated, apparently when the phenomena in question could not be
seen. On the purpose of the Diaries see Hunger and Pingree 1999 : 140 , contraBrown 2000 :
97 – 103 ; 2001. Data pertaining to eclipses that occurred as early as the mid-eighth century
BCare listed separately on cuneiform tablets, as are data pertaining to the location of the
planets and their phenomena, the earliest of which date to the Neo-Assyrian period. See now
Hunger 2001 and Huber and de Meis 2004.
17 See Hunger and Pingree 1999 : 189 for a possible seventh-century scheme comparable to one
particular column of the lunar ephemerides.
18 There is no definitive evidence that a distinction was ever made in Babylonian sources between
the equinoctial or sidereal years, a prerequisite to discovering precession (in some form).
Hipparchus is credited with this discovery. It was theorised that the Babylonians had indeed
discovered it first, but Neugebauer refuted this idea in the 1950 s (references in Hunger and
Pingree 1999 : 201 ). The fact that cuneiform texts linked the phases of a star (Sirius) with
seasonal phenomena does seem to support this (Neugebauer 1975 : 543 ), but Britton 2002 :
51 – 2 has raised the possibility that values of the sidereal and equinoctial years were indeed
differentiated in late cuneiform sources.
19 Meaning that the longitude of the syzygy is determined using a step function which has
longitude as its argument. System B lunar ephemerides describe the longitude of syzygy with
a zig-zag function with month number as an argument. This is less consistent than system A.
20 Proposed in Brown 2000 : 166 , and now attested directly in BM 47494 : rev. 16 , edited by
Hunger 2004.
— David Brown —