19-year cycle in the form of two tables, thefirst containing a full lunar
calendar, and the second (derived from thefirst) containing the dates of Easter
Sunday, through the 19 years of the cycle. In both tables, all dates arefixed and
given according to the Julian calendar; the lunar calendar and sequence of
lunar months is largely modelled, in fact, on the structure of the Julian
calendar. Nevertheless, Anatolius’cycle is discrepant by a few days from the
Julian calendar, as it assumes fewer leap years than the Julian calendar actually
has.^84 This distortion of the Julian calendar means that in practice, his cycle
would have been close to useless. It was a numerically ingenious attempt to
combine and harmonize different calendrical values, but highly unlikely ever
to have been used (Mc Carthy and Breen 2003: 142–3).
Some time later, a completely different 19-year Easter cycle was designed in
Alexandria, this one structured on the basis of the Alexandrian calendar (see
Chapter 5) and considerably more accurate in relation to both the lunar
month and the Alexandrian calendar.^85 The origins of this cycle are unclear,
but it seems to have been well established in Alexandria by the mid-fourth
centuryCE(more on this below). It soon became the norm in the Orthodox
Churches of the Roman East, and from the mid-fifth century was also adopted,
with modifications, in theWest.
The rise of Easter cycles in the later third-century East was obviously linked
to the slightly earlier emergence of Easter cycles in Rome and theWest;
however, Anatolius and other Alexandrians expressed different motivations
from those of pseudo-Cyprian, who suggested, as we have seen above, that
computation was intended for Christians to determine the date of Easter
independently from the Jews. For Anatolius and the Alexandrians, the pur-
pose of Easter cycles was to determine the true dates of Passover as Jesus
but over-speculative, as it is based on minimal evidence (Eusebius’citation) and some unproved
assumptions (such as continuity between Anatolius’and the later Alexandrian cycles).
(^84) Anatolius’cycle assumes a Julian calendar with only two bissextile (‘leap’) years in the
entire 19-year cycle, whereas in the Julian calendar there is one bissextile year every four years.
This deviation from the Julian calendar was necessary for Anatolius to obtain a cycle with afixed
number of days and for this number to be a multiple of seven—without which the Julian dates of
Easter Sundays could not have beenfixed and recurrent every 19 years (Mc Carthy and Breen
2003: 99 85 – 100). The tables are inDeRatione Paschali,10–11, ll. 159–97.
E. Schwartz (1905) 3–29, Blackburn and Holford-Strevens (1999) 803–5. The accuracy of
this cycle, in contrast to Anatolius’, derives from its greaterflexibility. The onlyfixed dates within
the Alexandrian cycle are those ofluna XIV, but the dates of Easter Sunday are variable from one
cycle to the next. It is often assumed that the Alexandrian cycle was an adaptation of Anatolius’
(e.g. Mosshammer 2008), but the new edition of Anatolius’work indicates in fact that apart from
being of 19 years, these cycles differed in every single point. Note, in particular, that Anatolius
provides a full lunar calendar (for all the months of the year, through the entire cycle), whereas
the Alexandrian cycle provides only a small number of lunar dates, most importantlyluna XIV
(Ethiopian computists, who have preserved the later Alexandrian 532-year cycle, do provide full
lunar calendar tables—see Neugebauer 1979: 79– 80 —but their antiquity is questionable).Who-
ever designed the Alexandrian cycle may have drawn some inspiration from Anatolius, but it is
best to regard it as an entirely new creation.
392 Calendars in Antiquity