Nevertheless, Athanasius’ability to determine the dates of Easter well in
advance—at least before the beginning of the 40-day fast of Lent preceding
Easter, of which the dates are also announced in the Letters—suggests at least
that some calculation or cycle was used. This is also implicit in the ability of
the Council of Serdica, in 343CE, to calculate the dates of Easterfifty years in
advance (more on this below). Furthermore, the sequence of Athanasius’
Easter dates in the Festal Letters until 348CE, and from 328 to 373CEin the
Index (the reliability of which, in this context, need not be doubted), confirms
that a 19-year cycle was followed, which points very clearly, though not
definitively, in the direction of the Alexandrian cycle.^117
The situation in early fourth-century Rome is even less clear. The 84-year
cycle of thesupputatioRomana—which, as discussed in Chapter 6 (near
n. 94), begins in 298CEand may have been designed already in the late third
century—is onlyfirst attested in the Calendar Codex of 354CE, where it was
evidently used for compiling the list of epacts which runs from the foundation
of Rome until 354CE(section VIII of the Codex).^118 But the list of dates of
Easter which runs from 312CEin section IX of the Codex does not conform
consistently with thesupputatio: thus in 333CEthesupputatiowould have had
Easter on 25 March, but the Codex has 15 April, and in 340CEthesupputatio
has 6 April, but the Codex 30 March;^119 more generally, thesupputatioallows
Easter to occur as early as 22 March, whereas the Codex and other evidence
suggests that the rule in Rome for most of the fourth century was not to allow
‘seventeen’and which necessitates, in that case, an epact of 30. But for 342CE, the manuscript of
the Index reads as lunar date‘sixteen’, clearly an error and emended by Schwartz to‘ 20 ’so as to
agree (in this case) with an epact of 30; yet in the Syriac or in the original Greek text,‘sixteen’is
far more likely to have been an error for‘nineteen’than for‘twenty’, which would corroborate
the epact of 29 that is explicitly given in our text. The Index, with its latersaltus, possibly reflects
small changes that the Alexandrian cycle underwent in the course of the 4th–early 5th cc. (so
Mosshammer 2008: 178–82). Indeed, one Easter table in a 7th-c. Latin manuscript (MS Paris Lat.
10318) agrees with the data of the Index, and may thus be preserving the cycle that was used by it
(Mosshammer 2008: 151–2, 181–2).
(^117) This does not help to establish when the Alexandrian cycle as later known might have
been instituted. Richard (1974) 310–15 assumed that it was instituted in 322/3CE, thefirst year of
an Alexandrian 19-year cycle; others (see Lejbowicz 2006: 45) have suggested the beginning of
the previous cycle, 303/4CE, when Peter of Alexandria was involved in a debate about the date of
Easter (on which see Stern 2001: 66–7, 72). Both these dates, however, might be too early (see
previous n.). See also E. Schwartz (1905) 3–29, Blackburn and Holford-Strevens (1999) 803–5.
We should not assume, in any event, that the Alexandrian cycle (or any other cycle) was
necessarily instituted itsfirst year (cf. Bickerman 1968: 41–2). As to the authorship of the
cycle, Jones wavers between Athanasius himself in 328CE(1943: 24) and Eusebius (1943: 26
n. 2), who according to Jerome (De Viris Illustribus61) designed a 19-year Easter cycle at some
time in this period (C.W. Jones 1943: 24 n. 2; Mc Carthy and Breen 2003: 137–8); but as
Jones recognizes, the authorship of the Alexandrian cycle remains completely speculative.
(^118) See Ch. 6 n. 94. The epact in Roman Easter cycles is the lunar date of 1 January.
(^119) However, the latter could be explained, following E. Schwartz (1905) 57–8, as a deliberate
deviation from thesupputatiofor the sake of compromising with Alexandria, as will be explained
below (in 340, according thesupputatio, 30 March wasluna XV: see table ibid. 47–8).
404 Calendars in Antiquity