Nevertheless, there are early traditions that suggest an association of the
beginning of the Egyptian lunar month with invisibility of the old moon. The
hieroglyph forAbd, the second day of the lunar month, includes a moon
crescent.^52 This makes sense in the context of a month beginning atfirst
invisibility, which means that the new moon crescent would normally become
visible on day 2 in the evening.^53 This inferencefinds support in a Middle
Kingdom (early second millenniumBCE) coffin text which states that the moon
(Thoth) is small atAbd, and great at the feast ofsmdt(half-month).‘Small’,in
opposition to‘great’, i.e. the full moon at the half-month, presumably refers to
a new moon crescent at itsfirst sighting; since this occurs on day 2 of the
month, the month presumably begins atfirst invisibility. Clearer still is a later,
Ptolemaic inscription from Karnak, where Khons (god of the moon) is
described as‘conceived on the feast ofpsd
ntyw(day 1 of the month), born
on the feast ofAbd(day 2), and mature on the feast ofsmdt(half-month, day
15)’.^54 This confirms that the period when the moon becomes invisible, or
when the new moon is‘conceived’, corresponded to day 1; whereas the new
moon’s birth, when it becomesfirst visible, was day 2.
This last text raises the question, however, of whether the lunar month
should have begun more appropriately at the time of the new moon’s‘concep-
tion’or at its time of‘birth’. In some Egyptian lunar calendars, indeed, the
latter appears to have been favoured. Thus the temple service months—i.e.
monthly periods allocated in rotation to different priestly groups (or‘phylai’)
for the performance of various temple duties—were lunar but counted from
the feast ofAbd, i.e. from the second day of ordinary lunar months.^55 This
one-day discrepancy between lunar and temple service months introduces a
further complication in our understanding of the Egyptian lunar calendar. It
suggests perhaps that the feast ofpsd
ntyw, when the moon is invisible, was
regarded as an ambiguous, interstitial day that could serve both as the last day
of the old lunar month and asfirst day of the new.^56 Above all, it shows that
lunar calendars were not uniform in Egypt but reckoned in a variety of ways.
(^52) Depuydt (2009) 117 n. 2, e.g. in a 19th-c. document from Illahun (pap. Berlin 10006, in Luft
1992: 39–42), and in an inscription from the temple of Medinet Habu, early 12th c.BCE(Clagett
198953 – 99: ii. 268).
Visibility of the new moon on day 1 in the evening would have been possible but rare; on
day 3 it would have been less rare, especially if weather conditions were unfavourable. See Stern
(2001) 99–100.
(^54) Parker (1950) 12; Clagett (1989–99) ii. 280–6; but see Clagett’s reservations ibid. 22),
pointing out that evidence of this kind (implying that the lunar month began on the day of
invisibility of the old moon) is generally late. 55
Luft (1992) 205–8 on the Illahun archive, early second millenniumBCE; Bennett (2008) and
Lippert (2009) on the Ptolemaic and Roman periods. Temple service months are usually called
wrš, whereas lunar months areAbd, which suggests a formal distinction between them. But not
all the evidence is conclusive: e.g. in pap. Berlin 10056 (see further below) it cannot really be
determined whether the service months that are listed begin onpsd
ntywor onAbd, even though
these months are clearly lunar (see Luft 1992: 189–95, Depuydt 1997: 147–8).
(^56) Similarly to the termwp-rnpt(see above and below, nn. 32, 69): Leo Depuydt, pers. comm.
144 Calendars in Antiquity