Sothis, whereas the new lunar calendar was regulated by the civil calendar,
with thefirst lunar month always following the civil New Year. According to
Depuydt (2009), this new lunar calendar came about and superseded the old
calendar in about the fourteenth centuryBCE, when the dates of the rising of
Sothis and the civil New Year were concurrent, and when the latter took the
place of the former asfixed reference point of the lunar year. Indeed, the
synchronic relationship between the lunar and the civil calendars becomes
evident in subsequent centuries, when month-names were commonly trans-
ferred between the two calendars.^71
The clearest attestation, however, of the new lunar calendar is a much later
document, pap. Carlsberg 9, which Parker took as evidence that in the fourth
centuryBCEempirical observation of the moon was abandoned in favour of a
fixed, cyclical scheme. This document, discovered andfirst published in 1938,
presents a comprehensive 25-year lunar cycle with nine intercalated years. It
assumes that 25 years of the Egyptian civil calendar equal exactly 309 lunar
months (which is a reasonable approximation); consequently, the lunar
months in this 25-year cycle recur on the same Egyptian civil dates, which
pap. Carlsberg 9 extensively lists. This lunar calendar is thus not regulated by
the rising of Sothis, but tied completely to the civil calendar; just like the civil
calendar, it recedes by one day in four years in relation to the seasonal (or
Sothic) year.^72
The document itself, pap. Carlsberg 9, is relatively late and could not have
been produced before the mid-second centuryCE, because it mentions a
25-year cycle beginning in year 7 of the Roman emperor Antoninus (144CE)
and an earlier cycle beginning in year 6 of emperor Tiberius (19CE). Parker
argues, however, that this lunar calendar must have been conceived and
instituted already in the fourth centuryBCE. It is only in this period, indeed,
that the dates in this document would have conformed to the day offirst
invisibility of the old moon, when the Egyptian lunar month traditionally
began. More precisely, in 357BCE(Parker’s preferred date for the institution of
the cycle), thefirst 25-year cycle would have begun on the day offirst
invisibility of the old moon (equivalent to 20 December), and likewise the
majority of lunar months in the cycle. In subsequent centuries, however,
because of a slight discrepancy in the 25-year cycle,^73 the dates of the lunar
months (and more particularly, of thefirst month of the cycle) would have
(^71) Depuydt (2009); on the transfer of month-names see also id. (1997) 161–7.
(^72) Parker (1950); Clagett (1989–99) ii. 23–4, 295–306.
(^73) After 25 years, the mean conjunction of the moon recurs on the same date of the Egyptian
civil calendar, but 1^1 / 5 hour earlier. This small discrepancy only becomes calendrically significant
after 20 cycles (500 years), when it accumulates to one whole day: the astronomical lunar month
then falls behind the civil calendar (and hence also behind the 25-year cycle, since the latter is
based on the civil calendar) by one day.
150 Calendars in Antiquity