Athenian origin is evident in their normal use of Athenian month-names,
although recent discoveries reveal that the month-names of other civil calen-
dars could also sometimes be used.^86
An eight-year cycle called octaeteris, comprising three intercalated years (of
13 months), is attributed by Roman-period writers to Greek astronomers from
as early as the sixth or earlyfifth centuriesBCE. A 19-year cycle (comprising
seven intercalated years) was devised in the latefifth centuryBCE; it is variously
attributed to the Athenian astronomer Meton (hence its common name
‘Metonic’) or one of his contemporaries.^87 This cycle is very accurate and
would have maintained the lunar calendar in an almost completely stable
relationship with the solar year. The Metonic cycle seems to have consisted, at
least by the early fourth century, of a full sequence of named months.^88 It was
later refined by the Athenian astronomer Callippus with a 76-year cycle
(consisting of four 19-year Metonic cycles minus one day). All the months
in this longer cycle were given specific, pre-determined lengths, which resul-
ted in a fullyfledged calendar (rather than just a sequence of months).
Thefirst Callippic cycle began in the summer of 330BCE.^89 Callippic dates
were widely used by Hellenistic astronomers, because they offered a stable,
(^86) The month-names of a Corinthian-type calendar have recently been identified in the
Metonic calendar represented on the Antikythera mechanism ofc.100BCE, possibly pointing
to a Syracusan origin: Freeth 87 et al.(2008) 614–16, (supp. nn.) 14–17. See further below, n. 93.
The creation of the Metonic cycle, or rather perhaps its epoch, is dated to the summer
solstice of 432BCE: see above, n. 16, Samuel (1972) 44–6, and Lehoux (2005). On Meton and the
relationship between this cycle and its Babylonian equivalent, see Bowen and Goldstein (1988)
and Hannah (2005) 54 88 – 7.
The evidence is in Ptolemy’sAlmagest(4. 11), and consists of three dates of Babylonian
lunar eclipse observations from 383– 2 BCEthat are given according to the Egyptian calendar and
an Athenian month. The Athenian months are unlikely to have been drawn from the civil
(festival) calendar, because by the time these reports of eclipses arrived in Greece and their
original Babylonian dates were converted into Egyptian and Athenian dates, past dates in the
Athenian festival calendar would have been difficult to recall and anyway of limited significance
(Toomer 1984: 12, 211–13). The month of the third report is called‘Poseideon I’(i.e. Poseideon
was repeated that year), which indicates that these names are not mere translations of Babylo-
nian month-names (not least because the equivalent Babylonian month, presumably Kislimu,
would never have been repeated in the Babylonian calendar), but rather belong to an autono-
mous calendar. Since the Callippic calendar was not yet in existence, the most likely candidate
might be a Metonic cycle (according to A. Jones 2007: 165 n. 3, this hypothesis is‘plausible
enough but cannot be confirmed’). It should be emphasized—as this is often overlooked—that
these Athenian month dates (which, significantly, do not specify the day of the month) only
prove the existence of a sequence of named months with an indication of which months in
the cycle are repeated (hence intercalary), but not—as in the Callippic cycle—a sequence of
months with pre-determined month-lengths; it would not necessarily have been, therefore,
a fully-fledged calendar as the Callippic was later to be.
(^89) The Callippic calendar (alongside earlier cycles) is described in some detail by Geminus,
Elem. Astr.8. 26–60 (see further below). See Ginzel (1911) ii. 409–19; Pedersen (1974) 127–8;
Samuel (1972) 47–9; A. Jones (1997) 157–8 and most importantly (2000a); Lehoux (2007) 90–4.
50 Calendars in Antiquity