The calendar of the temple

The weekday funnel of the previous chapter closed on one point: given a Wednesday, only AD 31 remains in the whole field. But that funnel hung on an assumption I then deliberately deferred, and which must now come onto the table — for it is the narrowest link in the whole case. The Wednesday of 25 April exists only if the Sanhedrin inserted a thirteenth month in the spring of 31, a second Adar. Without that intercalary month Passover falls a month earlier, on Tuesday 27 March, and the Wednesday is gone. This chapter digs that one assumption out, down to the bottom — how the calendar of the second temple worked, how probable an Adar II in 31 was, what it costs, and what the sharpest opponents themselves concede about it.

I state the nature of the outcome in advance, for it is modest and honest. This chapter delivers no proof that the intercalary month fell — no one can, and it does not pretend to. It delivers a maximum: how strong this assumption can be made, and where its limit lies. And it shows that the question is symmetrical, which is often forgotten. "14 Nisan = 27 March" is likewise not a fact but an assumption; it presupposes that the Sanhedrin inserted no Adar II. Both parties fill in the same missing decision. The difference is that the one does it aloud and the other in silence.

Begin with the misunderstanding that must be cleared from the table. The calendar of the second temple was no printed almanac, no table to be looked up in advance. It was alive. A month began not on a computed day, but on the evening two witnesses saw the first crescent and confirmed it before the council. A year received its thirteenth month not according to a fixed cycle, but when the barley was not yet ripe and the council decided: not yet, we wait a month. Bread, moon, barley, a meeting of men — on these it turned. The fixed, computed nineteen-year cycle we know, that of Hillel II, did not yet exist; tradition dates it around 359, a good three centuries later.

Those two rules stand, both of them, in a single chapter of an ancient source — and it pays to read them precisely, for they decide more than one might think. Eusebius preserves in his church history the Paschal canon of Anatolius of Alexandria, and there they lie side by side. The one rule, Anatolius' own, anchors the month: the first month may not be placed before the equinox; whoever does so "errs not a little". The other rule, which Anatolius reports from the ancients — Philo, Josephus, and the teachers of Aristobulus — anchors the feast: the paschal lamb must be slain "after the spring equinox, at the middle of the first month" (Eusebius, Historia Ecclesiastica 7.32.14–18). Two rules, each with its own bearer. And Anatolius shows that it is one doctrine seen from two sides: if the month lies right, the fourteenth falls of itself with the sun in Aries and the full moon opposite it. They diverge only in the tail — and AD 31 lies precisely in that tail. One honesty belongs here explicitly, and it cuts both ways: the Aristobulus citation behind rule B reaches us only at third hand — Anatolius quotes the ancients, and Aristobulus' own text is lost — so that rule B is more weakly attested than Anatolius' own rule A.

That honesty must reach further than the Aristobulus citation, for it touches the whole list of witnesses. Anatolius names Philo, Josephus and the teachers of Aristobulus as the ancients behind rule B. Whoever checks those names does not find the rule with them.

Philo names the equinox in six places, and in every one of them it is a season. He calls the two equinox feasts "many-dayed" (πολυήμεροι) and says that each of them lasts seven days; within a single equinoctial period he places the first, the tenth and the fifteenth of the month; and he writes that each of the two equinoxes takes place "in a seventh month". One day cannot be all three at once. Philo explains by means of the equinox why a month is called the first — he ties no feast day to it (De opificio mundi 116; De Decalogo 161; De specialibus legibus I 181–186 and II 150–152, 204; De vita Mosis II 222–224, all read in the Greek text of Cohn and Wendland, with the critical apparatus).

And Josephus nowhere names the equinox; the word has not been found anywhere in his work. What he gives is that the fourteenth falls while the sun passes through Aries, and that is a description: no "one must" stands beside it, it is the only astronomical placement in his whole oeuvre, and it stands as an addition to a verse of the law that does not have it (Antiquitates 3.248; Leviticus 23:5).

Anatolius' list of witnesses is therefore an attribution, not a citation. Sacha Stern says as much of both names — the attribution of the rule "to Philo and Josephus, whose works we are well acquainted with, is actually debatable", and the rule itself "is not explicitly enunciated in any of their works" — and he extends the doubt to the third name with a motive attached: the attribution to ancient Jewish authorities "suited the needs of Anatolius", who was justifying his own Easter computation, "and we are entitled, therefore, to question the impartiality of his interpretation" (Stern, Calendar and Community, pp. 51, 61).

And now the measure that this lays on me. Whoever downgrades a witness because he attributes freely may not forget it the moment that same witness suits him. Rule A is Anatolius' own rule, and rule A is the rule that gives the Wednesday in AD 31. If he handles loosely what the ancients taught, why should his own rule stand?

The answer is that rule A does not rest on his authority. It rests on a practice that is measurable apart from him: in the Babylonian tables Nisan begins after the equinox, year upon year — "the Babylonian Nisanu always began after the vernal equinox", writes the same Stern — and that can be counted without opening a single church father (Parker & Dubberstein, 1946, p. 46; Stern, Calendar and Community, p. 61; appendix A.3). What Anatolius describes there can be checked independently of him; his report about the ancients cannot.

That is the difference, and it is the only thing that counts here. A rule that can be read off the practice survives the weakness of its spokesman. An attribution does not.

For the main point of this chapter nothing changes; the count that follows stands as it stands. What changes is the reason why rule A carries weight — not because Anatolius states it, but because the tables show it — the Babylonian practice, that is; whether Jerusalem kept it is another question, and it follows.

One counter-practice belongs beside it. In Egypt the Jewish Passover kept to the equinox: Aristobulus of Alexandria — known only through Anatolius' report, see above — has it fall, in the mid-second century BC, at or around the equinox, "which implies on average one month before the Babylonian 14th/15th Nisan", and Stern adds: "The same is apparent in Philo's works" (Stern, 2000, p. 171 n. 69) — in the seasonal sense set out above; Philo ties no day to it. Stern also grants that where crops and fruits were ripe a Passover before the equinox "may have happened quite frequently" (2001, p. 70; ch. 7). There are thus two practices in the record, not one. The Judean dates that can be checked against external evidence, and the Berenice inscription, follow the Babylonian months and fall late (Stern, 2001, p. 61); the Egyptian practice falls a month earlier. Rule A does not decide which of the two Jerusalem kept. It rests on the Judean data, which are few, and that choice stands in this book as an assumption, not as a finding (appendix D). [DOCUMENTED — Stern, ZPE 130 (2000) 171 n. 69, seen at first hand]

Work both rules through for AD 31, and something emerges that sums up the whole case of this book in one sentence. Under neither of the two ancient equinox rules does a Friday exist in AD 31. Take Anatolius' rule, which anchors the month: then the winter March moon (nine days before the equinox) is rejected, Nisan falls in April, and 14 Nisan lands on Wednesday 25 April. Take the strict rule Anatolius reports from the ancients, which demands the feast after the equinox: then the March moon stands, and 14 Nisan falls on Tuesday 27 March. Wednesday or Tuesday — but under neither a Friday. Whoever wishes to carry the Friday consensus for AD 31 can appeal to neither of the two ancient rules. Under every ancient reading the year falls outside the Friday. That is the definitive, source-borne form of what Chapter 1 stated: the exclusion of AD 31 is not a measurement, but an artefact of the axiom.

But which of the two options for AD 31 — the April Wednesday or the March Tuesday — is the real one? The equinox rule does not decide that; the intercalation does, and it points consistently to April. For the council inserted a month on the strength of three signs: the aviv (the ripeness of the barley), the fruit of the trees, and the tequfah (the state of the season); if two of the three lagged behind, one intercalated (Tosefta, Sanhedrin 2:2, Machon Mamre edition; b. Sanhedrin 11a–b). And there is a preserved letter showing precisely this decision, dictated from the steps of the temple mount, in Aramaic: "the young doves are tender and the lambs are thin, and the time of the aviv has not come; [...] and I have added thirty days to this year" (b. Sanhedrin 11a–b). Two archaic marks — the temple steps, hence before the year 70, and the Aramaic letter formula — argue for high antiquity. Precisely the grounds that could shift a year by a full month.

And here is a precision the debate often misses. The bare 14 Nisan of 27 March itself falls after the equinox (22/23 March). The tequfah ground — "Passover after the equinox" — was thus already satisfied in the early option; it forced no intercalary month. Whoever wants an Adar II in 31 cannot appeal to the equinox; the decision must have rested on the barley and the tree fruits, the two agricultural signs. Were those, in the cold spring of 31, so late that the council added a month? That is beyond proof — but it can be calibrated.

The field itself has been surveyed, and it fixes two things no table gives. The first is the threshold. The field authority Gustaf Dalman notes in his Arbeit und Sitte in Palästina that the "aviv" is not harvest ripeness but Weichreife — the dough stage, in which "the grains of at least the barley are developed, though not yet fully ripe" (AuS I/2, p. 450), the stage that "precedes the threshed harvest on the threshing floor" (AuS III, p. 10). Aviv thus lies about one to two weeks before the harvest, not on it; and four witnesses point the same way — Dalman twice, Leviticus 2:14 (the roasted ear: to roast, there must be a grain) and Mishnah Menachot 10:9 (the omer must come "from the moist"). The second is the regional spread, and it is large enough to decide. For the council weighed not one field but the ripeness in two of three regions — Judea, the Transjordan and Galilee (Tosefta, Sanhedrin 2:2, Machon Mamre edition). ⭐ And that regional rule does not rest on a single transmission: the Bavli gives it as an independent baraita, in the same two-of-three form — "תנו רבנן: על שלש ארצות מעברין את השנה — יהודה, ועבר הירדן, והגליל. על שתים מהן מעברין, ועל אחת מהן אין מעברין" (b. Sanhedrin 11b). [DOCUMENTED — 11a–b examined at first hand in the Aramaic, 5 Sept. 2026]

Two anchors, apart from AD 31 altogether, show that the April placement is not plucked from the air. The first is AD 30: there 14 Nisan falls robustly on Friday 7 April, under every observing rule — a check that the computational method is sound (ch. 7). That addition is no hedge but a necessity: under the later, computed molad calendar AD 30 would in fact give a Wednesday, and the same reckoning takes the Wednesday away from AD 31 as well (ch. 14). It is thus the observing premise that orders the whole field, and it holds here as well as there. The second is stronger and seldom used. Six years after the crucifixion, in AD 37, the legate Vitellius is in Jerusalem for what Josephus calls "the ancestral feast" (ἑορτῆς πατρίου — almost certainly Passover, but the identification is a construal and not a verbatim designation), and there, "on the fourth day", learns of the death of the emperor Tiberius, who died on 16 March (Josephus, Antiquitates 18.89–90, 120–124). The travel time of the death report makes a March Passover physically impossible; the feast of AD 37 must have lain in April — and the Babylonian table gives precisely that: 14 Nisan AD 37 = 19 April (Parker & Dubberstein, 1946, p. 46). A documented late-April Passover, outside the whole crucifixion question, confirming that the observational calendar really could place the fourteenth of Nisan that late. Late Passovers existed.

And now the bounding of that second anchor, for it is needed and I have measured it myself. The year AD 37 confirms that a late-April Passover was possible; it does not discriminate between the reconstructions that diverge on this point. Richard Lanser's Exodus calendar, which rejects the April moon of AD 31 and places it a month earlier, arrives for AD 37 at precisely the same 19 April — because AD 37 in his scheme does not follow a Babylonian leap year, and there the two calendars coincide again. Whoever sets the Vitellius year against that reconstruction, then, sets it wrongly: it confirms that one just as well as this book's. It carries what stands above — that the observing calendar could place the fourteenth of Nisan that late — and not a millimetre more. [DOCUMENTED; the bounding CALCULATED]

A third anchor carries where the second leaves off — provided one states honestly what it carries, and what not. It does not decide the dispute over the intercalation: Parker too applies the Babylonian series to Jerusalem, and thereby fills in the same undocumented decision of the Sanhedrin that every reconstruction here must fill in. What it does do is this: it is a computation for Jerusalem instead of for Babylon, made by an author who had no interest whatever in a Wednesday — he wrote, on the contrary, against a reconstruction that wanted to reach the Friday. And it yields the April placement of AD 31 independently of this book. That is less than a decision and more than an echo. In 1944 Richard A. Parker — the Parker of the Babylonian tables — published in the Journal of Biblical Literature a refutation of Grace Amadon's calendar theory, and he opens it with a table he computes expressly for Jerusalem: "from the standpoint of the first possible visibility of the crescent at Jerusalem", with Schoch's new-moon tables and a longitude correction of thirty-seven minutes. His results for the seven years that matter:

Nisan 1 on 12 April gives 14 Nisan on 25 April — and that is a Wednesday. [DOCUMENTED — Parker, 1944, pp. 173–174, read at first hand; the weekday CALCULATED, calibrated on JDN 1732497 = Friday 27 April AD 31]

Three things make this weightier than it looks. First, Parker's seven dates agree to the day with the Babylonian table — he computed for Jerusalem, P&D for Babylon, and the difference is nil; the opposition "Babylonian versus Jerusalem" is thus empty for these years, as far as the astronomy goes; which intercalation practice Jerusalem kept is the separate question above. Second, Parker's standard is purely observational: he rejects Amadon precisely because she must set Nisan 1 twice on the third day of visibility — "It is inconceivable to the writer that any ancient people using a lunar calendar should follow such a practice". And third, and this weighs heaviest: Parker wrote against an Adventist chronology. He had no Wednesday to defend and no interest in AD 31. The April moon of that year is thus not the outcome of a theological choice; it is the outcome of the standard computation, published by the man who made the tables.

And then the star witness, for the case turns. The most-cited astronomical defence of a Friday in AD 33 is that of Humphreys and Waddington — and precisely they, at the decisive point, give the argument away. They concede that there are "no historical records" of intercalary months in the years 26–36, that Nisan "in some years" may have lain "a month later", and that they tested that late scenario on Fridays only (Humphreys & Waddington, 1992, p. 336). But within their own scenario, left open, precisely Wednesday 25 April appears for AD 31 — and their own eclipse table computes that very day: "25 April AD 31 — Wednesday — 35%", a partial lunar eclipse, which occurs only at full moon. They do not reject that Wednesday. They never consider it. The strongest opponent supplies, unwittingly, the weekday and the full moon of the model date.

Where does the inquiry stand, then, at the end of the calendar flank? At an honest maximum, and that is more than the assumption at first seemed able to carry. Wednesday 25 April 31 is astronomically exact and calendar-historically possible. Under Anatolius' own rule it is the forced outcome; under no ancient rule is there a Friday that year; the intercalation rests on attested, primary grounds; an intercalary year was a necessity of the system; an equally late Passover stands in black and white six years later; and the sharpest opponent himself computed the Wednesday and the full moon of that evening. It stands or falls with two assumptions: that Jerusalem kept the late, Babylonian practice and not the Egyptian one, and — unprovable but well-founded — that the barley of one cold spring lagged; both are assumptions the opponents themselves leave open. The calendar flank is thereby not demonstrated, but made as strong as it can be made, and no longer a concealed assumption but an excavated, testable position.

Astronomy has now given its candidates, and the calendar has said which month it allows. There remains the question of which years the career of Jesus allows — the baptism, the ministry, the years of Tiberius and the temple. That is the next question.

Note on the sources

The astronomy (new moon 11 March 31; equinox 22/23 March; 1 Nisan 12 April; 14 Nisan with Adar II = Wed 25 April, without = Tue 27 March; AD 37 = 19 April) can be verified and stands in appendix A; the text of Eusebius (HE 7.32) has been checked in the Greek (Loeb), and the two equinox rules have each been assigned to their own bearer (rule A is Anatolius' own; rule B is his report of the ancients). The six Philo passages on the equinox were read on the page image of the Greek of Cohn and Wendland, each with its critical apparatus; the word "equinox" has not been found in Josephus and the Aries clause of Ant. 3.248 is a description, so that Anatolius' list of witnesses is treated here as an attribution and not as a citation — with Stern (pp. 51, 61), who says the same of Philo and Josephus. Stern's note on the Egyptian practice (ZPE 130, 2000, p. 171 n. 69) has been read at first hand; it makes the Babylonian conformity of Jerusalem an assumption, not a finding. The rabbinic intercalation texts (Tosefta Sanhedrin 2:2, Machon Mamre edition, with the rule of "two of three regions") were consulted at first hand; the barley field corpus — Dalman, Arbeit und Sitte in Palästina I/2 and III, and the regional figures of Bauer that he passes on — has been seen at first hand from the scanned editions and carries the aviv threshold (Weichreife) and the regional spread, while the modern dataset of Strashny (2017) stands beside it as confirmation, not as ground; the sun-in-Aries passage has been weighed as a category, not as a measurement — on the page image of Niese I, p. 207, with apparatus and testimonia, and on a count of ten dating formulae in Josephus, in none of which the temporal genitive absolute lays a condition on the day. The solar longitude of 25 April AD 31 (≈ 32.4°, 2.4° of Taurus) stands in appendix A.1b and is entered there as a cost. This chapter does not prove the intercalary month; it makes it as strong as the data allow, and shows that AD 31 knows no Friday under any ancient equinox rule. Nowhere is a proof claimed here.

Literature (APA)

Bauer, L. (1902). Volksleben im Lande der Bibel. American Tract Society. (Barley-harvest regional figures: p. 126.)

Beckwith, R. T. (1996/2001). Calendar and chronology, Jewish and Christian. Brill. [pp. 282–283 seen primarily]

Dalman, G. (1928). Arbeit und Sitte in Palästina, Band I/2: Der Jahreslauf — Frühling und Sommer. C. Bertelsmann.

Dalman, G. (1933). Arbeit und Sitte in Palästina, Band III: Von der Ernte zum Mehl. C. Bertelsmann.

Eusebius of Caesarea. (1932). The ecclesiastical history (K. Lake & J. E. L. Oulton, Trans.; Book 7, ch. 32) (Loeb Classical Library). Harvard University Press.

Humphreys, C. J., & Waddington, W. G. (1992). The Jewish calendar, a lunar eclipse and the date of Christ's crucifixion. Tyndale Bulletin, 43(2), 331–351. (p. 336)

Josephus, F. (1887). Flavii Iosephi opera, vol. I: Antiquitatum Iudaicarum libri I–V (B. Niese, Ed.). Weidmann. (Ant. 3.248 on p. 207; the page image with critical apparatus and the testimonia line — "§ 248 = Leuit. XXIII 5" — seen at first hand. No variant touches ἐν κριῷ τοῦ ἡλίου καθεστῶτος; witnesses ROMSPLLat.)

Lanser, R. D. (2026). Reconstructing the original Hebrew calendar. Associates for Biblical Research.

Milgrom, J. (2001). Leviticus 23–27 (Anchor Yale Bible, Vol. 3B, p. 2062). Doubleday.

Parker, R. A. (1944). Ancient Jewish calendation: A criticism. Journal of Biblical Literature, 63(2), 173–176. (Nisan 1 for AD 28–34, computed on the first visibility of the crescent at Jerusalem; AD 31 = 12 April. Read at first hand.)

Parker, R. A., & Dubberstein, W. H. (1946). Babylonian chronology 626 B.C.–A.D. 45 (Studies in Ancient Oriental Civilization 24, p. 46). University of Chicago Press. (Also consulted in the machine-readable edition of the tables, pp. 25–46, in which Nisanu 1 of AD 31 stands on 12 April, JDN 1732482, and the Ululu II of AD 30 on 18 September, JDN 1732276.)

Philo of Alexandria. (1896–1906). Philonis Alexandrini opera quae supersunt (L. Cohn & P. Wendland, Eds.). Georg Reimer. (Greek with critical apparatus; used: vol. 1, p. 41 = De opificio mundi 116; vol. 4, p. 304 = De Decalogo 161 and p. 252 = De vita Mosis II 222–224; vol. 5, pp. 44–45 = De specialibus legibus I 181–186, p. 122 = II 150–152 and p. 137 = II 204. Page images seen at first hand.)

Stern, S. (2000). The Babylonian calendar at Elephantine. Zeitschrift für Papyrologie und Epigraphik, 130, 159–171. [p. 171 n. 69: the Egyptian practice, Aristobulus and Philo.]

Stern, S. (2001). Calendar and community: A history of the Jewish calendar, 2nd century BCE–10th century CE (pp. 61, 70, 102–103, 115, 124). Oxford University Press.

Strashny, A. (2017). Modern searches for Aviv barley in the context of the Hebrew calendar: A first description of the Israeli barley observation data. Jewish Bible Quarterly, 45(3), 179–187.

Primary texts: Tosefta, Sanhedrin 2:2 (Machon Mamre edition) · Babylonian Talmud, Sanhedrin 11a–b, 12a · Babylonian Talmud, Menachot 85a · Mishnah, Rosh ha-Shanah 1:7; 2:6–8 · Mishnah, Menachot 10:2; 10:8; 10:9 · Mishnah, Pesachim 4:8 · Leviticus 2:14 · Josephus, Antiquitates Judaicae 3.248; 18.89–90, 120–124 · Philo, De opificio mundi 116 · De Decalogo 161 · De specialibus legibus 2.26; I 181–186; II 150–152, 204 · De vita Mosis II 222–224. Greek/Aramaic via the editions named; my own working translation. Astronomy: own reproduction (ephem/Swiss Ephemeris), calibrated; appendix A. For Hezekiah: 2 Chronicles 30:2–3, 15 · Numbers 9:10–11 · Mishnah, Rosh ha-Shanah 2:8–9 · Mishnah, Pesachim 4:9 · Babylonian Talmud, Berakhot 10b (Sefaria, checked at first hand).