epigraphic evidence, of the regularity of Athenian intercalation and its possible
conformity tofixed cycles remains entirely to be made.^64
Strepsiades and the count of daysOne further passage in Aristophanes’Clouds(1131–4) has been taken to
suggest that already in thefifth centuryBCEthe length of Athenian months
was regular and predictable—even if this would run counter to another
passage in the same comedy, which has been cited above, where the moon
complains of calendar disruption. Because of its importance, detailed attention
will be given to this passage here. In this passage, Strepsiades counts down the
days until the end of the month, from thefifth day to‘old and new’(the last
day of the month) which he is dreading.^65 Strepsiades’ apparently exact
knowledge of how many days are left in the month suggests, atfirst sight, a
calendar set in advance without the possibility of irregularity or tampering.^66
In actual fact, however, Strepsiades’countdown is quite unremarkable, as it
represents the way Athenian days of month were normally reckoned. The last
ten days (‘decad’) of the Athenian month were reckoned backwards from
‘10th’to‘1st’(‘1st’being the last day, normally called‘old and new’), as was the
practice in many other Greek calendars.^67 This backward count appears to
(^64) Even more problematic is the evidence of Athenian‘new silver’coinage, which some
scholars in the 1960s attempted to use as evidence of a 19-year cycle (notably Meritt 1964; see
also Bickerman 1968: 35, 100 n. 33, Samuel 1972: 59). It has since been shown that these coins are
insecurely dated, and indeed cannot be dated independently from the epigraphic evidence;
consequently, they cannot prove anything that is not already known from inscriptions
(Mattingly 1971: 39–43, Müller 1991).
(^65) ‘Dayfive, day four, day three, after that day two, then the day that above all days intimidates
me...because the next day is the Old and New Day, when every single one of my creditors has
vowed tofile a lawsuit against me’(Henderson 1998: 161–3; commentary in Dover 1968: 231).
According toWalsh (1981) 112 n. 10 and Dunn (1998) 217 n. 16,‘old and new’was a debt-
collection day.
(^66) And hence, possibly, that a regular alternation of full and defective months could be
assumed (as suggested by myself in Stern 2001: 103 n. 17). Alternatively, however, Strepsiades
could have been relying on an earlier, monthly announcement of the month length by the
archon. Many scholars have used this passage in support of the theory of the gibbous moon (see
below, nn. 68, 70, 73). 67
Woodhead (1992) 141 n. 28,contraSamuel (1972) 59–61. On decads in Greek calendars,
see above, n. 13. De Blois (1998) 16, referring to South Arabian calendars that appear to have
used a similar system, suggests, as a rationale for this backward count, that the intention was for
the moon crescent to be the same size in the same-number days of thefirst and last decad (e.g.
3rd and 28th of the month, the latter also called‘3rd’). However, even in the most accurate lunar
calendar this correspondence is unlikely to have obtained. In particular, if day 1 of the month
followed thefirst appearance of the new moon (as assumed in the Greek calendars), on day 30
(and probably also day 29) the moon would have been invisible; which is sufficient to demon-
strate that thefirst and last decad could not have been symmetrical with regard to the appearance
and size of the moon crescent.
44 Calendars in Antiquity