In contrast to the Old Persian calendar, the Persian Zoroastrian calendar
was non-lunar andfixed. It consisted of twelve months of 30 days, plusfive
additional (‘epagomenal’) days—thus in structure, the same as the Egyptian
civil calendar. In fact, its months were conterminous with Egyptian months.
Both calendars were therefore identical, with the only differences that the
Persianfirst month, Farwardīn, corresponded to the Egyptian fourth month
(IV Akhet or Choiak), and that the epagomenal days were positioned else-
where in the year (more on this later).^26 This almost complete identity could
not possibly have been fortuitous, and leads to the inescapable conclusion that
the Persian Zoroastrian calendar was directly modelled on the Egyptian.^27
Like the Egyptian civil calendar, the Persian 365-day calendar drifted from
the seasons by one day every four years. Various Muslim authors from the
tenth centuryCE(culminating with al-Biruni, inc.1000CE) acknowledge that
this held true in their own period, but claim that in the pre-Islamic period
(before the seventh centuryCE) the Persians corrected this drift by intercalat-
ing one 30-day month every 120 years.^28 But the historicity of this claim has
been rejected by de Blois (1996) as a mere legend or unsubstantiated theory,
on the grounds that one-month intercalations are not mentioned in
any contemporary or earlier Zoroastrian source (although the notion of
intercalation is discussed, rather unclearly, in Zoroastrian sources from the
ninth–twelfth centuriesCE) or in any other source from Antiquity.^29 Positive
evidence that the calendar was not intercalated in the pre-Islamic period (or
later) could be inferred, I would suggest, from the date given in various sources
for Ardashir’s victory over the Parthian king Artabanus IV in 224CE.^30 But a
(^26) The month-names of the Persian Zoroastrian calendar are Farwardīn, Urdībihišt, Xurdāä,
Tīr, Murdāä,Šahrīwar, Mihr,Ābān,Āäar, Day, Bahman, Isfandārmuä; the epagomenal days are
called the Gatha days.
(^27) As already recognized in the 16th c. by J. J. Scaliger, in his seminal work on calendars and
chronology,De Emendatione Temporum(1583): see Hartner 1985 (756), de Blois 1996 (48–9).
Bickerman’s suggestion (1968: 43) that the Persian Zoroastrian calendar was based on the
Babylonian‘business year’of twelve 30-day months (on which see Ch. 2, near nn. 54–6) is
plainly wrong.
(^28) According to some, the frequency was every 116 years; the sources are also inconsistent
about the dates of these pre-Islamic intercalations (de Blois loc. cit.).
(^29) Boyce (2005) 13–19 rejects de Blois’s argument and maintains the theory of intercalation in
the Sasanian period, but her argument is uncritical of the sources and considerably less
convincing. 30
According to the SyriacChronicle of Arbela, ch. 8, the date wasWednesday. 27 Nisan 535SE
(i.e. 224CE), which may be corrected to 28 Nisan (= 28 April), as 27 Nisan was a Tuesday (Sachau
1915: 61 and n. 1). According to the Arabic chronicle of al-T:abari, on the other hand, it was the
last day of (the Persian month) Mihr (Nöldeke 1879: 14; Bosworth 1999: 13–14 and n. 54). The
dates of 28 Nisan and 30 Mihr coincide in this year only if we assume that the Persian
Zoroastrian calendar was not intercalated (if we correct instead the weekday in theChronicle
of Arbela, and thus prefer the date of Tuesday 27 Nisan, we still have approximate compatibility
to a non-intercalated calendar, whereas an intercalated calendar remains completely incompati-
ble). More ambiguous is the date of the martyrdom of Anahid, which is given in the Syriac
Martyr Acts (Bedjan 1890–7: ii. 603) as‘in the (Syriac) month of H:aziran, on day 18, that is the
TheRise of the Fixed Calendars 175