Printer Friendly

The date of death of Jesus of Nazareth.

BORN INTO A WORKING-CLASS JEWISH HOME at Nazareth in Galilee a couple of years before 1 C.E., Jesus early on devotes himself to rabbinical learning. After leading a life of anonymity until about age thirty, he sets out to preach his message. His followers consider him the "Anointed One," Messiah in Hebrew, Christ in Greek. But some in the establishment interpret his success as a threat and conspire to eliminate him. One Thursday evening in the month of Nisan, in early spring, after sharing a last meal with his disciples, Jesus is arrested. The next day, a Friday, he is tried and put to death. In the following centuries, the religion founded on his teachings spreads throughout the inhabited world.


The day Jesus of Nazareth died--the day of the Crucifixion or Passion in Christian tradition--occupies a special place in chronology. No topic has been examined as thoroughly. There are two levels at which one may engage the topic. One can study it as a rite of passage: the serious study of chronology presupposes familiarity with its great problems. Or one can try to say something new. One type of new statement is impossible--proposing a new date. Only a handful of days enter into consideration. Everyone has always agreed that the day was (1) a Friday, (2) falling close to full moon, (3) in early spring, (4) in the late 20s or early 30s of the first century C.E. This paper's aim is to restore the year date 29 C.E. , once generally favored and then entirely discarded, to its original status. It is the year best supported by the sources.

By the early l900s, 29 C.E. had come to be generally accepted as the year of Jesus's death. 18 March 29 C.E. is reported in volume 3 of the well-known eleventh edition of the Encyclopaedia Britannica (Turner 1910). It seemed as if the last word had been said on the matter. But new developments around the turn of the century, notably the decipherment of Babylonian astronomy and the discovery of Aramaic double dates in papyri from Egypt, provided new impetus to astronomical chronology. As a result, 29 C.E. was dismissed, the reason being astronomical. If Jesus died in that year, then lunar Nisan, the month in which he died, had to have begun before the first crescent could have been seen. The first crescent is initially seen above the western horizon in the evening soon after sunset one to two days after new moon. New moon is when the moon is directly between the sun and the earth and therefore invisible from the earth. One to two hours after first visibility, the first crescent sets below the horizon, followi ng the sun.

The hunt was on for an alternative date. In 1930, in the distinguished science journal Astronomische Nachrichten, O. Gerhardt concluded (1930: 162) from an "objective" examination of all the evidence that, "clearly," Jesus died on 7 April 30 C.E. Still, he hoped (p. 139) that readers would find the materials presented with sufficient transparency to judge for themselves and might even reach different conclusions. A different conclusion followed soon enough. In 1934, the eminent J. K. Fotheringham (1934: 161), classicist turned astronomer, argued for another date, 3 April 33 C.E., because it "on the whole.., offers fewer difficulties than any of the others." But he added, "[M]y ambition has been rather to explain the character and tendencies of the different lines of evidence than to arrive at a conclusion." The fountain of ingenuity seems then to have been exhausted. Decades passed until two recent contributions in ZDMG by W. Hinz (see section 4 below) reopened the issue.

The argument raised against 29 C.E. will not be disputed here. My aim is rather to show that the assumption on which that argument is founded is not in the least binding.

Command over all that is written on this topic would be difficult to achieve. But no such command seems necessary to produce a new line of argument. What follows is a source study. Should the sources somehow have it completely wrong, we cannot move beyond them. Then there are also the patent contradictions in the sources. The most striking is that some sources place the death of Jesus on 14 Nisan; others, on 15 Nisan. If absence of contradiction were required, all efforts at dating the event would need to be abandoned. But the contradictions also mean that not all sources count. At least some must be wrong. To some extent, historical inquiry is less about establishing what happened than about preparing and critiquing sources so that others might judge in their own right what occurred.

The divine status of Jesus in the Christian tradition has sometimes made it difficult to examine his life fully by the dispassionate criteria of the historical method. Thus, the valuable light shed on the topic by Jewish sources has only recently come to be fully appreciated. Two milestones are Morris Goldstein's outstanding Jesus in the Jewish Tradition (1950) and Joseph Klausner's earlier Jesus of Nazareth (1925), the latter translated from the Hebrew by Herbert Danby.

The present paper has four parts. The first two concern the year of the event. In part one, the evidence in favor of 29 C.E. is adduced. Its reliability is then scrutinized. In part two, the main objection to 29 C.E. is first described and then exposed as being the result of an unsubstantiated assumption. Part three concerns the month and day of the event. In 29 C.E., the choice is between 18 March and 15 April, the only two Fridays falling close to full moon in early spring that year. 18 March is preferred here. Part four analyzes the most recent contribution to the topic. I shall make every effort to render the argument accessible and transparent to those with no prior knowledge of the topic.

1. 29 C.E.

1.1. The Evidence

According to respected Latin church authors dating from the second century C.E. onward, the year of Jesus's death was that in which the two Gemini were Roman consuls, 29 C.E. (1) It seems fair to characterize 29 C.E. as the "traditional date" transmitted by ancient Christian sources. The gospels give no clear indication of the year of the event. Much effort has been expended to extract a year date from them, with different and uncertain results.

The full tripartite names of the two consuls, consisting of praenomen, nomen, and cognomen, are Gaius Rufius Geminus (Paulys Real-Encyclopadie VII [1912]: 208) and Lucius Rubellius Geminus (ibid. IA.l [1914]: 1160). The Roman year, of which the modern calendar is a continuation, already ran from i January to 31 December at the time. The order and the identities of the consuls of Rome are so well known that it seems superfluous to adduce bibliographical references. A readily accessible list of the consuls for each year from 509 B.C.E. to 337 C.E. is Bickerman's (1980: 140-62; for 29 C.E., see 154). It is based on comprehensive publications of the lists by A. Degrassi and T. R. Broughton.

At least six considerations diminish the chance of error in the transmission of the consul date. First, dating by consuls was normal in Rome; rare dating methods might be more prone to error in transmission. Second, a pair of personal names is more distinctive and memorable as a year date than a mere year number. Third, the fact that the two consuls share the cognomen Geminus is distinctive and memorable. Fourth, there are no consuls of that cognomen in the decades preceding or following 29 C.E. Fifth, Rome ruled Palestine when Jesus died. The event was therefore dated directly in the Roman calendar. Conversions of dates from one calendar to another present more opportunities for error. Sixth, Pontius Pilate, the Roman administrator who sentenced Jesus, later returned to Rome; and Peter and Paul were martyred there. These connections bridge the distance between Palestine in the east, where the event happened, and the Latin sources in the west transmitting the year date.

1.2. Is 29 C.E. True?

All depends on how truth is defined. Each element in the present argument ought to be defined. That includes the concept of truth. In my view, truth is an impression of the individual mind. Thus, it cannot exist independently of the minds of people contemplating it. Two distinct impressions of the mind may be singled out. The first is the inability to answer the question, How else could it have been? The second impression is absolute freedom from that same question-so much so that one would be utterly baffled if the impression were contradicted. These two impressions differ in quality. Both are perceived as that which is often called truth.

Instances of the second impression are the dates of the major events of the twentieth century C.E. We would be dumbfounded if it appeared that they had happened at dates other than those reported in the history books. Likewise, we would be speechless if we learned that we had been born in another year than we had thought we were. This second impression may be called the sense of absolute truth. It is ultimately a perception of the mind. Indeed, it is occasionally contradicted. It is difficult to produce the sense of absolute truth regarding the consul dating. The second impression, that of absolute truth, is passive. The mind does not seek alternatives. By contrast, the first impression is dynamic; the mind is searching. There are two main types of results: inability or ability to produce alternatives. Inability is close to the second impression here called absolute truth.

In the present case, alternatives are conceivable. For example, the event may have been falsely associated with the year of the two Gemini at a later time. The earliest report of the consul date is two centuries younger than the event. The gospels mention only one year date in relation to the public life of Jesus, namely year 15 of the emperor Tiberius. That year is mentioned at Luke 3:1 in connection with the beginning of the public life of John the Baptist. By some theories, only about a year separated that event from the Friday of the Crucifixion. The length of the ministries of John and Jesus is just one of the related topics that have been discussed tirelessly with no definitive results. It is not clear how the years of Tiberius are counted in the gospels, or in most other sources for that matter. The emperor Augustus, Tiberius's predecessor, died on 19 August 14 C.E. Counting from that day, Year 15 would strictly have reached from 20 August 28 C.E. to 19 August 29 C.E. But this is just one conceivable way of counting.

The consul date might therefore result from efforts by later church authors to interpret Luke's year 15. Year 15 of Tiberius is sufficiently close to the consulate of the two Gemini by any year count that someone might have chosen that consul year from a list of consuls. Tacitus begins Book Four of his Annals (IV.1.1) with the date C. Asinio C. Antistlo consulibus nonus Tiberio annus. The consulate of 23 C.E. is equated with year 9 of Tiberius. The consul year begins on i January. It is not clear whether Tacitus thought of year 9 of Tiberius as beginning on that same day. Owing to his political convictions, Tacitus otherwise never mentions regnal years of emperors (Koestermann 1965: 31). But if the consulate of 23 C.E. could be associated by an eminent historian like Tacitus with year 9, then it is easy to see how the consulate of 29 C.E. might have been equated with year 15 (9 + 6). Once the association found its way into an authoritative source, its further transmission would be difficult to halt.

The surviving sources do not permit the positive confirmation or rejection of this alternative scenario. The mere ability to produce an alternative diminishes the value of the consul date as historical evidence. Such is the nature of historical inquiry. The state of the sources imposes permanent limitations. But this handicap does not relieve historians of the task of calibrating the veracity of items reported in the sources. The final result is often at best a kind of open-ended calibration that the true scientist may not find satisfying. Items of historical evidence are not, after all, subject to experiments reproducible in a laboratory.

But this paper's focus is not on the consul date. Rather, it is on a point of astronomical chronology that has wiped 29 C.E. off the map of modern chronology.

1.3. Clement and Eusebius

Nothing in the sources rivals the consul date in clarity. But two items from Clement (b. Ca. 150 C.E.) and Eusebius (ca. 260-340 C.E.) may be adduced, even if only to illustrate the nature of the other evidence.

In Book 1 of his Stromateis, at 145, 5 (ed. O. Stahlin), Clement states that forty-two years and three months separate the fall of Jerusalem (which happened in the summer of 70 C.E.) from the Passion. Origen (ca. 185/86-ca. 254/55) mentions the same forty-two years, but with some reservation, in Against Celsus, at IV.22. In his Homily on Jeremiah, at XIV.13, Origen refers to the same forty-two years more assertively as the time separating year 15 of Tiberius from the fall of Jerusalem. Apparently, the assumption is that Jesus died in year 15 of Tiberius. This year date is mentioned in Luke 3:1, but not in direct connection with the event (see 1.2).

On this basis, the resulting year would be 28 C.E. (70-42), not 29 C.E. Upon closer inspection, however, Clement's account (Strornateis 1.145, 2-5) may be seen to be internally inconsistent. Clement counts (1) about thirty years from Jesus's birth to baptism, (2) one year to his death, (3) forty-two years and three months to the fall of Jerusalem, and (4) 122 years, 10 months, and 13 days to the death of Commodus. But the total from the birth of Jesus to the death of Commodus is given as 194 years, 1 month, and 13 days. One would rather expect 196 years, I month, and 13 days. Something is wrong.

In his Ecclesiastical History (ed. E. Schwarz), at I.13.22, Eusebius (ca. 260-340 C.E.) cites a text translated from Syriac reporting certain legendary events as occurring soon after the Assumption of Christ. The year date is 340. The count is clearly that of the Seleucid Era (SE). In cuneiform documents, years SE are counted from the spring of 311 B.C.E.; in Syriac Christian sources, they are counted from the autumn of 312 B.C.E., half a year earlier. Hinz (1989: 306) identifies SE 340 with 28 C.E. But years SE do not fully overlap with Julian years. Furthermore, 340 SE begins in spring 29 C.E. in cuneiform documents but in fall 28 C.E. in Greek and Syriac sources. Neither includes the spring of 28 C.E. (2) The Syriac Christian count, which must be the one Eusebius uses, includes the spring of 29 C.E., and therefore agrees with the consul date.


2.1. Day 1 of Nisan and First-Crescent Visibility

2.1.0. The objection may be defined as follows. The sources inevitably lead to the conclusion that, if Jesus died in 29 C.E., then the lunar month of Nisan began before first-crescent visibility in that year.

2.1.1. First-Crescent Visibility

How soon after new moon or conjunction can the first crescent be seen? New moon can fall at any time of the day, but of course the first crescent is initially visible in the early evening. Much has been written on first-crescent visibility. Kepler thought the problem was unsolvable. In a sense, it has never been solved; there are too many variables. But more is known now than in Kepler's time. Thus, the world record for soonest naked-eye sighting of the crescent was recently still above fifteen hours after new moon (Schaefer, Ahmad, and Doggett 1993; Pepin 1996). Of course, it is not because astronomers tell us that the crescent could have been seen one evening that it indeed historically was. But at least, modern computation can determine by which evening the first crescent could not yet possibly have been seen. That is what we need for the present argument. Also, the possibility that people sometimes may think they see the crescent when it is not there cannot entirely be disregarded.

Little is known about ancient estimates of the distance from new moon to the evening of first-crescent visibility. By an ancient rule of thumb, the crescent cannot be seen within less than a day or twenty-four hours. This estimate is not all that inaccurate. Epping (1890: 49-53) first derived more precise ancient estimates from information found in astronomical cuneiform tablets kept at the British Museum and relating to the years 124 to 122 B.C.E. and 111 to 110 B.C.E. Epping's information is still fresh for the present purpose.

The tablets actually list times between sunset and moonset on the evening of first-crescent visibility. These times range between forty minutes and ninety minutes. From these times, one can derive the times between conjunction and sunset. They are all well above fifteen hours, the lowest being just under nineteen hours. The following numbers refer to [ minutes]: 45 [] 43 minutes [31.sup.52], [44.Sup.39], [35.sup.13], [26.sup.37], [41.sup.57], [32.sup.43], [22.sup.20], [34.sup.44], [37.sup.17], [30.sup.12], [37.sup.58], [45.sup.23], [29.sup.26], [27.sup.06], [41.sup.17], [33.sup.36], [26.sup.47], [44.sup.03], [50.sup.50], [32.sup.26], [38.sup.15], [37.sup.22], [25.sup.07], [35.sup.20], [44.sup.12], [52.sup.12], [40.sup.13], [28.sup.35], [18.sup.43], [34.sup.06], and [25.sup.33].

Pioneers like Epping (1890) and Kugler (1907-24, 1; 65) assumed that the time between sunset and moonset was computed rather than observed. Many now accept that some distances were observed because they are so precise (see Brack-Bernsen 1997: 99 on tablet Cambyses 400 or BM 33066). However, due to cloud cover, some first crescents cannot possibly have been observed. These were presumably calculated (on this, see now Brack-Bernsen 1997).

2.1.2. Day 1 of Nisan in 29 C.E. if Jesus Died in that Year

There are only four possibilities: 4 March or 5 March, 1 April or 2 April. These four dates for 1 Nisan are derived from two properties of the day Jesus died, according to all the sources--with the gospels leading the way. First, it was a Friday. Second, it fell near full moon close to the spring equinox.

The second property is obtained as follows. The gospels leave no doubt that Jesus died at the beginning of Passover. The Passover holiday begins in the late afternoon of 14 Nisan with a sacrifice, followed that same evening by the sacrificial meal. The next morning is daylight of 15 Nisan. That day is celebrated as a Sabbath. The holiday lasts seven days from 15 to 22 Nisan. In one sense, Passover begins late in the day on 14 Nisan; in another, it begins on 15 Nisan. Nisan is lunar and always begins around a new moon close to the spring equinox. Passover therefore also begins around a full moon close to the spring equinox. (3)

It is not clear, however, whether the Friday afternoon Jesus died was 14 Nisan (in which case he died around or shortly before the time of the Passover sacrifice) or 15 Nisan (that is, the first full day of Passover, a Sab bath day). Everyone has always agreed that the synoptic gospels, Matthew, Mark, and Luke, can easily be interpreted as saying that it was 15 Nisan and John as saying that it was 14 Nisan, even if no evangelist gives an explicit day number. Efforts to resolve the patent contradiction in the gospels have not been successful. Considering all the evidence, this writer joins many others in assuming that John, who was there when Jesus died, is probably right and Matthew, Mark, and Luke are probably wrong. In that view, Jesus died on 14 Nisan. The number 14 is the norm in early Jewish, Christian, and Manichaean literature. However, both 14 Nisan and 15 Nisan will be considered in what follows.

On the evening before his death, Jesus shared a last meal with his disciples, the Last Supper of Christian tradition. There is no doubt from the gospels that this meal was planned as--and hence in effect was--a Passover meal. But if John's 14 Nisan is correct, then Jesus celebrated the Passover meal a day too early. Incidentally, a slightly different question is whether the Passover meal was a Seder. But the Seder ritual as we know it today probably developed only after 70 C.E. (see Bokser 1987 [with bibliography]), so Jesus's meal could not be a Seder, strictly speaking. Thus if John's 14 Nisan is correct, the Last Supper differs in two ways from the modern Seder. First, as a Passover meal, it was a forerunner of the Seder. Second, it was celebrated a day too early. It appears, then, that Jesus died on a Friday around full moon close to the spring equinox.

Now, Friday and full moon are both cyclical events. There is no reason to assume that the seven-day cycle of the week was ever disrupted, and we therefore know on what days all the Fridays of history have fallen. The week originates in Jewish custom. It first spread by being adopted in cities throughout the Roman empire around the beginning of the common era. Josephus (first century C.E.) observes in his Contra Apionem (at 11.39): "There is no Greek or non-Greek city to which our tradition of celebrating the seventh day has not spread." Its Biblical origin contributed much to making the week fully dominant when Christianity became a state religion in the fourth century C.E. Likewise, the cycle of full moons can be wheeled backwards by computation. Each full moon is separated from the preceding and the following by intervals that are on average about 29.53059 days long. In 29 C.E., only two Fridays fell around full moon close to the spring equinox, 18 March and 15 April. (4) It follows that both 18 March and 1 5 April could be either 14 Nisan or 15 Nisan if Jesus died in that year.

The dates for 1 Nisan can be simply derived. There is some controversy about the beginning of the day in antiquity. To sidestep this controversy, which is irrelevant to the present argument, the following equivalences concern daylight only, from sunrise to sunset:

If 18 March = 15 Nisan, then 4 March = 1 Nisan.

If 18 March = 14 Nisan, then 5 March = 1 Nisan.

If 15 April = 15 Nisan, then 1 April = 1 Nisan.

If 15 April = 14 Nisan, then 2 April = 1 Nisan.

2.1.3. 1 Nisan Preceded First-Crescent Visibility in 29 C.E. if Jesus Died in that Year

The question next arises: if daylight of 1 Nisan coincided with daylight of 4 March, 5 March, 1 April, or 2 April in 29 C.E., could the first crescent have been sighted on the evenings before, that is, on 3 March, 4 March, 31 March, or 1 April?

The two relevant new moons or conjunctions are 4 March 3:37 a.m. and 2 April 8:06 p.m. (Goldstine [1973: 86] for Babylon minus 37 minutes for Jerusalem, confirmed as still adequate by P. Huber [personal communication]). Three of the four evenings--3 March, 31 March, and 1 April-patently precede new moon. There is no way the first crescent can be seen before new moon.

As for the fourth evening, 4 March, sunset occurred under 15 hours after the new moon of 3:37 a.m. on 4 March. Fifteen hours after conjunction is not enough time for the first crescent to become visible, for it is lower than the world record of first-crescent sighting in modern times. Sunset occurred a little before 6:00 p.m. because it was close to the spring equinox; moonset followed less than an hour later.

In sum, 29 C.E. "cannot be forced into agreement with astronomy" (Fotheringham 1934: 158-59) and is thereby "completely eliminated" (Gerhardt 1930: 154) as the year of death of Jesus if lunar months cannot begin before first-crescent visibility. 18 March 29 C.E. then becomes a "desperate hypothesis" (Fotheringham 1934: 159). Only Hinz (1989: 305--6, n. 9) states the inevitable explicitly: the consul date must then be an "error."

Can 29 C.E. be salvaged? Obviously only if first-crescent sighting did not mark the beginnings of lunar months.

2.2. First-Crescent Sighting and the Beginnings of Lunar Months in Antiquity

That lunar months typically could not begin before the first crescent was sighted is, on closer inspection, one of ancient chronology's most overrated and persistent assumptions. By the practice of first-crescent sighting, watchers look out for the first crescent in the evenings after the last crescent has disappeared in the morning. The last crescent is last seen above the eastern horizon in the early morning shortly before sunrise one to two days before new moon. Presumably, as soon as the first crescent was sighted, it was decided to call the daylight starting the next morning day 1.

The assumption that this practice was typical in antiquity is in large part supported by the fact that it is undeniable in two cases. These two cases are the two pillars on which the presumed dominance of first-crescent sighting rests. First is the Muslim religious calendar. Second is the calendar of Babylonian astronomical texts, which evidence a sophisticated intellectual tradition lasting from the eighth century B.C.E. to the first century C.E., perhaps the longest-lasting unified research project in history. Even if first-crescent visibility was computed and not observed in Babylon, crescent visibility is clearly the standard. In both cases, the practice of first-crescent sighting was strictly supervised by highly centralized authorities, councils of clerics, or families of scholars. In neither case can one speak of a calendar arising naturally from popular use.

Babylonian practice may have resulted from how data were collected. The Babylonians carefully measured six intervals of time, each lasting about one to two hours between horizon crossings of moon and sun around new moon and full moon. A. Sachs has called them the Lunar Six. Horizon crossings are empirically discrete, easy to time, and natural targets of observation. Conjunction itself is invisible. Timing it was the aim; the Lunar Six were a means.

One of the six intervals was the time between sunset and moonset on the evening of first-crescent visibility. In careful record-keeping, a clear principle regulating the position of lunar day 1 is desirable. It may have been decided to call the daylight period that follows the evening of the first member of the Lunar Six day 1, even if this rule did not quite reflect popular usage. Indeed, cuneiform texts report instances in which the first crescent was seen at times other than the evening preceding daylight of day 1 (Kugler 1907-24, 2:14; cf. Beaulieu 1993: 86 n. 39). Beyond these two cases, positive evidence for first-crescent sighting as a calendrical practice is difficult to find. Evidence for other practices is not. The following evidence from Egypt, Greece, and Rome could easily be supplemented.

Egyptian lunar months clearly began before first-crescent visibility, even if it is not clear exactly how the beginning of the months were determined (on the Egyptian lunar calendar, see Depuydt 1997). First-crescent visibility has also descended from its pedestal in Athenian and Greek calendrics (see Pritchett 1982). This seachange is confirmed by the lunar date 13 Skirophorion, which is implied in Meton's dating of the solstice of 432 B.C.E. (Depuydt 1996: 42). If Skirophorion began after the first crescent was sighted, then the solstice was observed at least two days after it actually happened. This is improbable. Furthermore, the lunar months of the 76-year Callippic cycles used by Greek astronomers also began before first-crescent visibility. The first cycle began in 330 B.C.E.

As regards Rome, a reliable passage in Macrobius (Saturnalia I.15; see Ideler 1825-26, 2: 39) reports that a certain high priest determined on the basis of the size of the first crescent whether the Nonne would be day 5 or day 7. What has escaped explicit recognition so far is that the length of the lunar month was determined about a month earlier, and by means other than first-crescent sighting.

2.3. An Untidy but Natural Alternative to First-Crescent Visibility

Strong support for the dominance of first-crescent sighting as a calendrical practice in antiquity seems to derive from the absence of a clear alternative. If not by first-crescent sighting, how was the beginning of lunar months marked? The search for a tidy alternative has never been satisfied. But must it be tidy? Most peoples of the world, from Polynesia to Alaska and from South America to Siberia, use or have used lunar calendars (Nilsson 1920). Hardly any of these calendars operate by rigorous rules. Why then must ancient lunar calendars have done so?

There is in fact hardly a need for a tidy and rigorous system. The average astronomical lunar month is about 29.53059 days long. All peoples using lunar calendars must sooner or later have realized that calendrical lunar months do not ever need to be shorter than twenty-nine full days or longer than thirty days. In alternating 29-day and 30-day months, it is easy to keep up with the lunar phases in the sky. These phases readily impress themselves on the senses. If the first crescent is at some point sighted too early, before twenty-nine days have passed, or too late, a couple days after lunar day 1, those deciding for village, town, or country can easily correct the problem by their particular choices of 29-day or 30-day months. By the most simple of manipulations, the calendrical lunar months of twenty-nine or thirty days can be kept in line roughly yet perpetually with the astronomical lunar month. In sum, an untidy system is not necessarily ineffective.

Watching for the crescent still fulfills a function in such a system. It serves as a check that the astronomical month and the calendrical month have not drifted too far apart. Crescent-watching is well attested in the sources (for Mesopotamia, see Wacholder and Weisberg 1976 and Beaulieu 1993). But the sources never explicitly state that the next daylight was then called day 1.

2.4. Evidence for Lunar Months Beginning before First-Crescent Visibility

2.4.0. If Jesus died in 29 C.E., then lunar Nisan began before first-crescent visibility. That inevitably follows from the sources. The question arises: how did lunar months begin in Jerusalem in the first century C.E., In the later first millennium C.E., the Hebrew calendar became fixed while remaining lunar (see 2.4.3). As a result, most information about earlier calendrical practices ceased to be transmitted in the sources. With respect to the calendar, "Israel has wiped out her religious past with a wet sponge," writes Amadon (1942: 278), quoting Chwolson (1908: 165).

The sources do not tell us how lunar months began in Jerusalem in the first century C.E. Simply accepting the calendrical practice of first-crescent sighting as a given is therefore not an option. One cannot reject the possibility that lunar months typically began before first-crescent visibility. Five sets of evidence suggest that elsewhere they did. Two sets from Egypt and Greece have already been adduced in 2.2. Three additional sets historically closer to first century C.E. Palestine follow below, namely double dates in Aramaic papyri from Egypt (2.4.1), the Talmud (2.4.2), and the fixed Jewish calendar (2.4.3).

The following logic is then acceptable: because Jesus died in 29 C.E., as the consul date suggests, lunar months could indeed begin before first-crescent sighting in first century C.E. Jerusalem.

2.4.1 Double Dates in Fifth Century B.C.E.

Aramaic Papyri from Egypt

Ancient sources abound with lunar dates, but it can seldom be established where day 1 fell in relation to conjunction. One may assume that it was close. But how close, whether a little early or a little late, is impossible to know. A rare exception is a set of double dates in Aramaic papyri from Egypt dating to the fifth century B.C.E., when Persia ruled Egypt. Most are from an island in the Nile near Assuan, which is called Elephantine in Greek. These papyri document the life of Jewish communities in Egypt. A global reassessment of the chronology of these texts remains desirable. The new comprehensive edition by Porten and Yardeni (1986--93) will greatly facilitate such a task. The focus of what follows is only on the relation of day I to conjunction.

The double dates equate an Egyptian civil date with a Jewish-Babylonian lunar date. The lunar date by itself cannot be related to conjunction. But both the Egyptian date and conjunction can be dated in years, months, and days B.C.E. The Egyptian date is thus linked to conjunction. And by being equated with the Egyptian date, so is the lunar date. Fotheringham (1908, 1911) was the first to note and the only one ever to defend the position that the lunar months of the double dates seem to begin a little before first-crescent visibility. Thus, in a papyrus that had not yet surfaced in Fotheringham's time, Kraeling no. 1, dated to year 14 of Artaxerxes (Kraeling 1953; Porten and Yardeni 1986-93, 2: 58-59, with foldout 13), lunar 20 Sivan is equated with Egyptian 25 Phamenoth. Daylight of 25 Phamenoth is daylight of 6 July 451 B.C.E. If daylight of lunar 20 Sivan is then also equated with 6 July 451 B.C.E., then daylight of lunar 1 Sivan is daylight of 17 June, as follows:
 July Sivan
 6 = 20
-5 -5
 1 = 15
 30 = 14
-13 -13
 17 = 1

The new moon or conjunction closest to 17 June is that of 1:59 p.m. on 16 June (Goldstine [1973]: 46 for Babylon, minus 47 minutes for Assuan). 1 Sivan is close to new moon. So are all the other beginnings of Jewish-Babylonian months derived from the double dates. It is this fact that proves that the Jewish-Babylonian months were lunar.

The evening preceding daylight of 17 June falls too soon after the new moon of 16 June 1:59 p.m. for the first crescent to have been visible. In other words, the lunar month begins a little too early for first-crescent sighting to have served as the marker of its beginning. And that is typical of the lunar months of the double dates, as Fotheringham first noted. But Fotheringham's observation was soon contested. Criticism was easier to make at the time because few double dates were known. Responding to Fotheringham, Ginzel (1906-14, 2 [1911]: 48, 52) reaffirmed the practice of first-crescent sighting as "certain" (sicher). In fact, Fotheringham himself, when writing later (1934) on the date of death of Jesus, accepted first-crescent sighting as a given for Jerusalem in the first century C.E. He does not refer to his earlier work on the Aramaic double dates from Egypt.

Additional double dates have since emerged in papyri published in 1953 by Kraeling. These lunar months too begin typically a little earlier than expected. The time is right, then, to breathe new life into Fotheringham's thesis.

Soon after Kraeling's publication, Horn and Wood (1954) subjected all the known double dates to a detailed study. They also noticed without referring to Fotheringham that the lunar months seem to begin a little too early. They admit (p. 19) that "there is the possibility that the Jews in Elephantine did not entirely rely on the observation of the new crescent to determine the beginning of the new month." Yet, Horn and Wood hold Onto first-crescent sighting as the marker of the beginning of the lunar month, doing so by means of an interesting theory (1954: 6). In their footsteps, Porten retains the traditional view that "the month was determined by visibility of the new crescent at sunset" (1990: 15-16). The theory may be analyzed into thirteen discrete steps for clarity. Kraeling no. 1 is adduced as an example.

(1) It is certain that daylight of 25 Phamenoth is daylight of 6 July. (2) It is certain that lunar 20 Sivan is equated with 25 Phamenoth. (3) Horn and Wood then assume that the lunar day lasted from evening to evening and the Egyptian day from morning to morning. Steps (4) through (11) then follow naturally from the assumption made in (3).

(4) Egyptian 25 Phamenoth lasted from 6 July morning to 7 July morning. (5) Lunar 20 Sivan overlaps with 25 Phamenoth and therefore lasted either from 5 July evening to 6 July evening or from 6 July evening to 7 July evening. (6) The lunar and Egyptian dates therefore overlap either from 6 July morning to 6 July evening or from 6 July evening to 7 July morning. (7) Daylight of 20 Sivan is then either 6 July or 7 July. (8) Daylight of 1 Sivan is then either 17 June or 18 June. (9) The first crescent therefore could mark the beginning of the lunar month only if daylight of 1 Sivan is 18 June. The closest new moon occurred at 1:59 p.m. on 16 June (see above). (10) Daylight of 20 Sivan then ought to be 7 July. (11) 20 Sivan then overlaps with 25 Phamenoth from 6 July evening to 7 July morning.

(12) Horn and Wood therefore assume that the letter in question must have been written in the period of overlap given in (11). They propose composition on the evening after sunset on 6 July. By the same theory, the morning before sunrise to 7 July is also possible. (13) Horn and Wood then also assume that all the letters whose lunar months seem to begin a little too early were written in the evening after sunset.

Parker (1955: 272) considers this solution "wholly unlikely." It is not clear whether he is rejecting the notion that letters can be written at night-at least, that is how Porten (1990: 21) understands Parker. However, it is easy to agree with Porten that letters can be written at night. Porten cites the Talmudic treatise Gittin, which refers to documents written by day [LANGUAGE NOT REPRODUCIBLE IN ASCII] and documents written at night [LANGUAGE NOT REPRODUCIBLE IN ASCII] (2.2 [Jerusalem Talmud]; 17b [Babylonian Talmud]). In a Greek papyrus letter from Egypt, a scribe writes [LANGUAGE NOT REPRODUCIBLE IN ASCII] [beta] or "when I was about to go to bed, I wrote two letters" (Wilcken 1912:50,9-11).

However, the real question is not whether a scribe could write a letter after sunset. Nor is the question when exactly the lunar day and the Egyptian day began. Rather, the key question is: did a scribe writing letters in a single sitting from late afternoon into the later evening change the lunar day date by one number at sunset while retaining the Egyptian date? This question is merely implied in Horn and Wood's theory. The sources neither confirm nor deny such a practice. An explicit statement that "At sunset I change the day date in my letters" belongs to the stuff that chronologers' dreams are made of.

Much has been written about the beginning of the day in antiquity. The usual starting-point is descriptions in classical sources of differences in the beginning of the day among various peoples. But it is generally agreed that these descriptions are often wrong. This whole topic deserves to be reexamined. The most penetrating study by far is still Bilfinger's (1888). Suffice it to state here that most people most of the time simply think of the day as beginning in the morning and ending in the evening. This is the so-called natural day. (5) Extensions to before sunrise and to after sunset are included; the nature of human activity organically connects these extensions with the daylight period to which they are adjacent. It is reasonable to assume that Aramaic scribes in the goings-about of daily life thought of the day in the same way. It therefore seems unlikely that dates would change at sunset. Horn and Wood's proposal (1954) implies such unnatural changes in date. However, their proposal even then does not produce the desired result. Two lunar months in the double dates still precede first-crescent visibility.

A preliminary investigation of the Aramaic double dates has produced the following results which I believe support Fotheringham's observation.

There are fourteen completely preserved double dates free of philological problems that might interfere with attempts at chronological interpretation. All relevant facts are listed in Table 1. The structures of the Egyptian and Jewish-Babylonian calendars are assumed to be known. In thirteen of the fourteen dates, the Jewish-Babylonian month begins close to conjunction. That suffices to confirm that month's lunar character. In no. 3, the month begins several days early for an unknown reason.

Regarding the relation of lunar day 1 to first-crescent visibility, the fourteen dates can be subdivided into four groups, (a) to (d). As to the computation of the evening of first-crescent visibility, the dates in Parker-Dubberstein (1961) are used as the standard. A different evaluation of the many variables involved in the computation of first-crescent visibility might result in slightly different results. Also, the act of watching for the first crescent is subject to many specific circumstances on which all information has been forever lost, such as place of observation and the weather.

(a) Five dates (nos. 1, 8, 9, 10, and 14 in the table) fit the first-crescent visibility requirement without it being necessary to accept with Horn and Wood both that the lunar day date in letters was increased by one at sunset (which seems improbable), and that the five documents were written after sunset (which cannot be verified). Horn and Wood would see no problem with assuming that these five documents were written during daylight.

(b) Six dates (nos. 2,4, 7, 11, 12, and 13) only allow first-crescent sighting if one does make the two aforementioned assumptions. Horn and Wood would assume that these six letters were written after sunset.

(c) Two dates (nos. 5 and 6) do not seem to allow first-crescent sighting under any assumptions.

(d) No. 3 is problematic and therefore better disregarded. The beginning of its lunar month is several days removed from conjunction.

Only one date (no. 1) falls later than the first possible evening of crescent visibility under any assumptions.

That is a tight fit if first-crescent sighting indeed marked the beginning of lunar months.

The table's columns contain the following information: (1) is the designation of the papyrus in Porten's recent edition (1986-93) and in Cowley's (1923) or Kraeling's (1953) earlier editions; Aime-Giron (1931) first published the M(emphis) S(hipyard) J(ournal). (2) = (3) is the date equation given in the text. (4) converts (2) = (3) to lunar day 1. (5) is daylight of the Egyptian date in (4). (5) is probably also daylight of the lunar date. But in flora and Wood's theory, daylight of the lunar date could fall a day later, that is, on the date provided in (6).

(7) is the time of conjunction (Goldstine [1973] for Babylon, minus 53 minutes for Memphis or 47 minutes for Assuan). Because daylight of the lunar date could fall on two different days (see [5] and [6]), two evenings need to be checked for first-crescent visibility. These evenings are given in (8) and (10) and evaluated in (9) and (11). (13) is the evening of first-crescent visibility in Parker and Dubberstein's (1961) tables. (9) and (11) are the distances in time from new moon to the evening of first-crescent visibility, taken uniformly as 6:00 p.m. This distance is called the translation period.

2.4.2. The Talmud

The material in the Talmud dates to about 200 B.C.E. - 500 C.E. The Mishna core was compiled around 200 C.E. The Gemara, or commentary on the Mishna, was written down in 200-500 C.E. (6) The treatises Rosh Hashanah, Sanhedrin, and Arakhin are most relevant to the present topic. Four observations (A) to (D) apply.

(A) Nowhere does the Talmud state precisely how months began. In fact, Rosh Hashanah (20b) speaks of [LANGUAGE NOT REPRODUCIBLE IN ASCII] "the secret of the calendar." Apparently, a calendar council whose deliberations were secret decided the question. Perhaps in part owing to this secrecy, the Talmud transmits different interpretations on calendrical points. In general, the Talmud often juxtaposes contrasting views. This procedure adds to its value as a source: everyone gets to talk.

(B) Nowhere in the Talmud is first-crescent visibility explicitly mentioned as a marker of the beginning of the month. Sighting the first crescent did play a key role, which was rendered obsolete by the fixed calendar in post-Talmudic times. Thus, the treatise Rosh Hashanah deals in detail with the interrogation of witnesses coming to testify on crescent sighting. Adequate testimony allowed the new crescent to be sanctified by being declared [LANGUAGE NOT REPRODUCIBLE IN ASCII] "sanctified!" Maimonides wrote a treatise on the matter, [LANGUAGE NOT REPRODUCIBLE IN ASCII] "Laws for the Sanctification of the New Moon." Handbooks on chronology assume that sanctification had a calendrical function, namely to identify day 1 of the month. But nowhere does the Talmud say this. To the contrary, Rosh Hashanah preserves a tradition by which Adar was always twenty-nine days long or [LANGUAGE NOT REPRODUCIBLE IN ASCII] "lacking (a thirtieth day), hollow." This leads to the following question (20a): why profane the Sabbath to testify about the crescent? It is already known that the month has twenty-nine days. The answer is telling: [LANGUAGE NOT REPRODUCIBLE IN ASCII] "because it is a religious duty to sanctify about the sighting." This directive is repeated for emphasis. Clearly, in this case sanctification is not a calendrical act. A related statement is [LANGUAGE NOT REPRODUCIBLE IN ASCII] "no contradiction: prolonging it (the month) (from 29 to 30 days) is one thing, sanctifying it is another" (20a). Arakhin (9b) states that "it is not obligatory to proclaim a new moon on the basis of having seen it." Apparently, sanctification did not even always require crescent sighting.

(C) The Talmud occasionally hints at procedures other than observation for determining the beginning of months (see Zuckermann 1882: passim; Steinsalz 1989: 280 bottom). (7) No clear picture emerges from the surviving sources. The following passages provide only hints. According to Rosh Hashanah (19b), Adar preceding Nisan always has twenty-nine days ([LANGUAGE NOT REPRODUCIBLE IN ASCII]). This is not first-crescent visibility. In an intriguing passage (Rosh Hashanah 20b), Mar Samuel (Shmuel) Yarkhinai, the great astronomer of the Babylonian school of Nehardea (220-250 C.E.), claims to be able to make a calendar for all the Diaspora. There is no reason to doubt that he could. However, Abba the father of Rabbi Simlai replies:


Does the master know what is taught in the secret of the calendar, (namely the rule of) born before noon or born after noon?

Shmuel answers: "I do not." Then Simlai again: "Since the master does not know this, there must also be other things the master does not know." Apparently, things were done differently in Palestine than in Babylonia. Simlai's criticism does not challenge Shmuel's competence in astronomy. Rather, it addresses halacha, or religious practice. "Birth" here means conjunction, the moment when the moon is right between sun and earth. Importantly, the rule "born before noon or born after noon" does not focus on crescent visibility.

(D) Lunar months could at least sometimes begin before first-crescent visibility. By a tradition mentioned above, Adar always had twenty-nine days. Astronomical lunar months are on average a little longer than 29.5 days. Calendrical lunar months will therefore be twentynine or thirty days long. If Adar is always twenty-nine days long, then it is on average too short. Nisan will therefore on average begin a little too early.

Then there is the rule "born before noon or born after noon." It survives as follows in the fixed calendar: if conjunction falls between midnight and noon, the next day is day I of the new month. If the moon is born between noon and midnight, the beginning of the month is postponed a day. By this system, lunar months typically begin before first-crescent visibility. The same may be assumed for the earlier Talmudic version of the rule.

It is not clear how conjunction was timed in the first century C.E. Greek and Babylonian astronomers were able to compute conjunction fairly accurately. No information survives as to whether astronomers in Jerusalem were capable of the same feat. Then again, a practiced observer of the sky may easily acquire a sense of how much the moon moves in relation to the sun every day and hence estimate in advance roughly when the moon is directly in front of the sun.

Ginzel (1906-14, 2: 80-8 1) reports that the Samaritans do not link the beginning of the month to first-crescent sighting. They use ancient rules transmitted from generation to generation similar to the rule "born before noon or born after noon." But instead of noon, 6:00 a.m. is used. Mahler (1916: 36) believes that the Samaritans preserve the ancient Jewish understanding of [LANGUAGE NOT REPRODUCIBLE IN ASCII] "between the two evenings" (see note 3). Why not also the rule just discussed?

2.4.3. The Hebrew Fixed Calendar

Like its ancient predecessor, the modern Hebrew calendar is lunar, but the beginnings of the months are no longer determined every month by observing the moon. Instead, they are fixed by the use of a precise value of the average lunar month, namely 29 days, 12 hours, 44 minutes, and 3 1/3 seconds. (8)

The introduction of the fixed calendar is traditionally dated to the fourth century C.E. But most recent students of the problem now date the institution to the late first millennium C.E. What matters here is that the months of the fixed lunar calendar begin on average before first-crescent visibility (Schwarz 1872: 58). That must have been the case already when the calendar was instituted, because the value of the average lunar month is so precise. It is reasonable to assume, then, that the fixed calendar perpetuated an earlier practice.

In the fixed calendar, the first crescent typically first appears on the evening that follows daylight of lunar day 1. In that sense, the first crescent can be said to first appear on lunar day I. Nothing in the sources contradicts the surmise that this may have been the intent of ancient Jewish calendrical practice.


The only Fridays on which Jesus could have died in 29 C.E. are 18 March and 15 April. 15 April has been preferred because Nisan's full moon fell with certainty in April in 29 C.E. in the Babylonian calendar (Parker-Dubberstein 1961: 46). Olmstead did not hesitate to apply the "miraculous precision" of the Babylonian calendar to the present problem (1942: 279). But 18 March cannot be dismissed, as the following four observations suggest.

First, the Jewish calendar indeed derives from the Babylonian calendar, but there is no evidence that Nisan always began with the same spring new moon in Jerusalem and in Babylon. In any year, there are two new moons in the spring that fall close to the equinox. In fact, modern Jewish lunar months begin on average earlier in relation to the spring equinox than Babylonian lunar months did. This may reflect an older practice.

Second, in early Christian authors, 25 March prevails (cf. Loi 1971). 25 March was a Friday in 29 C.E., but Friday 25 March 29 C.E. fell a week after full moon and is therefore out of the question. Epiphanius (Panarion L.l.7-8, ed. K. Holl) reports that the Quartodecimans, early Christians who ate the Passover meal with the Jews on the evening of 14 Nisan, celebrated the Passover on 25 March because they had read in the Acts of Pilate that Jesus died on that day. Then again, Epiphanius reports having seen copies of the Acts of Pilate in which the day is 18 March. The origin of this intriguing tradition is unknown.

Third, if one extends the fixed Jewish calendar, instituted in the late first millennium C.E., backward into the past, 14 Nisan falls exactly on 18 March in 29 C.E. I have not found this fact noted anywhere. Its value is limited because the precise relation of the later fixed calendar to the earlier observational calendar is unknown.

Fourth, March is the month almost always mentioned in the sources. The sources disagree on the day. But then lunar 14 Nisan falls on a different day every year in the Julian calendar. Any association of 14 Nisan with a specific day in that calendar therefore distorts historical reality.

As for the time of day, the synoptic Gospels (Matthew 27:46, Mark 15:34, and Luke 23:44) state that Jesus died at the ninth hour. Whether the hours are seasonal (that is, last one-twelfth of daylight) or equinoctial (that is, last sixty minutes) does not make much difference. One-twelfth of daylight is about sixty minutes around the spring equinox.

In sum, there is much that points to ca. 3:00 p.m. (Jerusalem local time) on 18 March 29 C.E. as the time of the event.


After a long lull in the debate, Hinz (1989, 1992) advanced a new theory. He culled much useful information from earlier treatments. But many elements of his theory are merely "non-impossible" and not positively verifiable, as is often the case in chronological research. What is more, the theory was altered fundamentally soon after publication. The year 28 C.E. was retained, but the day was changed from Friday 30 April (Hinz 1989) to Friday 27 April (Hinz 1992). How can day 30 and day 27 of a month both be Fridays? The two seem only three days apart.

First of all, 27 April 28 C.E. has always been regarded as a Thesday by historians, not a Friday. (9) To make that Tuesday into a Friday, Hinz adduces a theory proposed by Strauli (1991) by which the emperor Constantine in 321 C.E. made "a Sunday into a Thursday (einen Sonntag zum Donnerstag)" (Hinz 1992: 55). If my understanding of this theory is correct, Constantine decided one fine Sunday that that Sunday had to be a Thursday, and the next day therefore a Friday instead of the Monday that it would have been. This reform is not mentioned in any sources, just inferred from reports of Constantine's worship of Jupiter, after whom Thursday is named in Latin, viz. Iovis dies (jeudi in French), Iovis "of Jupiter" being the genitive of lupiter. By this theory, if day 30 of a certain month was a Sunday, then Constantine's reform would "make that Sunday into a Thursday." Anyone unaware of the reform would assume that day 23 of the month, seven days earlier, had also been a Thursday, whereas it had in actuality been a Sunday. Therefore, before 321 C.E., what historians now think of as Thursdays would all have been Sundays, Fridays would have been Mondays, and so on.

However, does Hinz really think it possible, as his theory implies, that the Jews would have disrupted the Sabbath cycle at any time in history, let alone in the early fourth century C.E., on the orders of a Roman emperor at a time when Christinaity did not yet dominate? Earlier, around 200 C.E., Tertullian writes in his Apologeticus, at 16, that the Sabbath fell on Saturday (dies Saturni). It would need to be a Thesday if (1) Constantine's reform had occurred and (2) if the Jews had not adopted the reform. The following logic applies, however: since (2') the Jews could hardly have adopted the reform, (1') the reform never happened.

Passages from church fathers are adduced in support of Constantine's reform. But this evidence seems moot because the theory exhibits an internal error even in its later version (Hinz 1992). Three assumptions are made: (1) Jesus died on a Friday afternoon, 27 April 28 C.E. (2) Daylight of 27 April 28 C.E. was daylight of 14 Nisan. (3) Nisan began after the first crescent had been sighted.

If daylight of 14 Nisan is daylight of 27 April, then daylight of 1 Nisan (14 - 13) is daylight of 14 April (27 - 13). If daylight of 1 Nisan is 14 April, then the first crescent was seen the previous evening, 13 April. But the closest conjunction occurred at 4:44 p.m. on 13 April 28 C.E. (Goldstine 1973: 86). On the evening of 13 April, the moon was just about in front of the sun. Sighting the crescent that evening was out of the question. The earliest possible evening of visibility was therefore 14 April (Hinz 1992: 15). But Hinz then concludes, "Das Passahfest am 15. Nisan fiel ... somit auf den 28. April." If the crescent is first seen in the evening of 14 April, then daylight of 1 Nisan is daylight of 15 April. Daylight of 15 Nisan (1 + 14) is then 29 April (15 + 14), and not 28 April. The preceding day, when Hinz believes Jesus died, must have been 28 April, and not 27 April. 28 C.E. cannot therefore be the year, by the theory's own explicit assumptions.
Table 1

Completely Preserved Double Dates and Their Relation to New Moon or

Abbreviations: B2, B3, C3 = numbers assigned in Porten's edition
(1986-93); Cow = Cowley 1923; epag. = epagomenal day; Kr = Kraeling
1953; MSJ = Memphis Shipyard Journal (document C3.8 in Porten 1986-93)

 (1) (2) (3) (4)
 Papyrus Egyptian Lunar (2) = (3)
 date date reduced to
 lunar Day 1

No. 1 C3.8 9 Choiak 24 Adar 16 Athyr
 MSJ Year 15 Year 14 = 1 Adar
No. 2 B2.1 28 Pachons 18 Elul 11 Pachons
 = Cow 5 Year 5 Xerxes = 1 Elul
No. 3 B3.1 4 Thoth 7 Kislev 3 epag.
 = Cow 10 Year 9 Artaxerxes I = 1 Kislev
No. 4 B3.2 25 Phamenoth 20 Sivan 6 Phamenoth
 = Kr 1 Year 14 Artaxerxes I = 1 Sivan
No. 5 B2.7 10 Mesore 2 Kislev 9 Mesore
 = Cow 13 Year 19 Artaxerxes I = 1 Kislev
No. 6 B2.8 19 Pachons 14 Ab 6 Pachons
 = Cow 14 Year 25 Artaxerxes I = 1 Ab
No. 7 B3.4 9 Payni 7 Elul 3 Payni
 = Kr 3 Year 28 Artaxerxes I = 1 Elul
No. 8 B3.5 25 Epeiph 25 Tishri 1 Epeiph
 = Kr 4 Year 31 Artaxerxes = 1 Tishri
No. 9 B3.6 7 Phamenoth 20 20 Sivan 18 Mecheir
 = Kr 5 Year 38 Artaxerxes I = 1 Sivan
No. 10 B3.9 22 Payni 6 Tishri 17 Payni
 = Kr 8 Year 8 Darius II = 1 Tishri
No. 11 B2.10 12 Thoth 3 Kislev 10 Thoth
 = Cow 25 Year 9 Darius II = 1 Kislev
No. 12 B2.11 9 Hathor 24 Shebat 16 Phaophi
 = Cow 28 Year 14 Darius II = 1 Shebat
No. 13 B3.10 29 Mesore 24 March 6 Mesore
 = Kr 9 Year 1 Artaxerxes II = 1 Marcheshvan
No. 14 B3.11 8 Choiak 20 Adar 19 Hathyr
 = Kr 10 Year 3 Artaxerxes II = 1 Adar

 (1) (5) (6) (7) (8)
 Papyrus Daylight of Latest New moon Evening of
 Egyptian date daylight of closest visibility
 in (4) lunar date to (5) needed
 in (4) for (6)

No. 1 C3.8 4 Mar 471 5 Mar 1 Mar 4 Mar
 2:41 p.m.
No. 2 B2.1 26 Aug 471 27 Aug 24 Aug 26 Aug
 = Cow 5 6:13 p.m.
No. 3 B3.1 12 Dec 457 13 Dec 16 Dec 12 Dec
 = Cow 10 8:06 a.m.
No. 4 B3.2 17 June 451 18 Jun 16 Jun 17 Jun
 = Kr 1 1:59 p.m.
No. 5 B2.7 16 Nov 446 17 Nov 16 Nov 16 Nov
 = Cow 13 5:17 a.m.
No. 6 B2.8 13 Aug 440 14 Aug 12 Aug 13 Aug
 = Cow 14 7:16 p.m.
No. 7 B3.4 8 Sep 437 9 Sep 7 Sep 8 Sep
 = Kr 3 12:02 p.m.
No. 8 B3.5 6 Oct 434 7 Oct 4 Oct 6 Oct
 = Kr 4 8:43 a.m.
No. 9 B3.6 24 May 427 25 May 22 May 24 May
 = Kr 5 4:58 a.m.
No. 10 B3.9 17 Sep 416 18 Sep 15 Sep 17 Sep
 = Kr 8 5:32 a.m.
No. 11 B2.10 14 Dec 416 15 Dec 12 Dec 14 Dec
 = Cow 25 11:52 p.m.
No. 12 B2.11 18 Jan 410 19 Jan 17 Jan 18 Jan
 = Cow 28 2:54 a.m.
No. 13 B3.10 2 Nov 404 3 Nov 1 Nov 2 Nov
 = Kr 9 10:41 a.m.
No. 14 B3.11 18 Feb 402 19 Feb 16 Feb 18 Feb
 = Kr 10 8:14 p.m.

 (1) (9) (10)
 Papyrus Evaluating Evening of
 (8) visibility
 for (5)

No. 1 C3.8 [75.sup.h]25m 3 Mar

No. 2 B2.1 [47.sup.h][47.sup.m] 25 Aug
 = Cow 5
No. 3 B3.1 visibility 11 Dec
 = Cow 10 impossible
No. 4 B3.2 [28.sup.h][01.sup.m] 16 Jun
 = Kr 1
No. 5 B2.7 visibility 15 Nov
 = Cow 13 impossible
No. 6 B2.8 translation 12 Aug
 = Cow 14 period small
No. 7 B3.4 [29.sup.h][58.sup.m] 7 Sep
 = Kr 3
No. 8 B3.5 [57.sup.h][07.sup.m] 5 Oct
 = Kr 4
No. 9 B3.6 [51.sup.h][02.sup.m] 23 May
 = Kr 5
No. 10 B3.9 [60.sup.h][28.sup.m] 16 Sep
 = Kr 8
No. 11 B2.10 [54.sup.h][08.sup.m] 13 Dec
 = Cow 25
No. 12 B2.11 [39.sup.h][06.sup.m] 17 Jan
 = Cow 28
No. 13 B3.10 [31.sup.h][19.sup.m] 1 Nov
 = Kr 9
No. 14 B3.11 [45.sup.h][46.sup.m] 17 Feb
 = Kr 10

 (1) (11) (12)
 Papyrus Evaluating Parker-
 (10) Dubberstein
 evening of

No. 1 C3.8 [51.sup.h][25.sup.m] 2 Mar

No. 2 B2.1 translation 26 Aug
 = Cow 5 period small
No. 3 B3.1 visibility 17 Dec
 = Cow 10 impossible
No. 4 B3.2 visibility 17 Jun
 = Kr 1 impossible
No. 5 B2.7 visibility 17 Nov
 = Cow 13 impossible
No. 6 B2.8 visibility 14 Aug
 = Cow 14 impossible
No. 7 B3.4 visibility 8 Sep
 = Kr 3 impossible
No. 8 B3.5 [33.sup.h][07.sup.m] 5 Oct
 = Kr 4
No. 9 B3.6 [27.sup.h][02.sup.m] 23 May
 = Kr 5
No. 10 B3.9 [36.sup.h][28.sup.m] 16 Sep
 = Kr 8
No. 11 B2.10 visibility 14 Dec
 = Cow 25 impossible
No. 12 B2.11 visibility 18 Jan
 = Cow 28 impossible
No. 13 B3.10 visibility 2 Nov
 = Kr 9 impossible
No. 14 B3.11 translation 17 Feb
 = Kr 10 period small

An abridged version of this paper was read at the 211th annual meeting of the American Oriental Society, in Toronto, on 2 April 2001, 9 Nisan 5761.

(1.) The two oldest are Tertullian (ca. 160-ca. 240), Against The Jews VIII.18 (sub Tiberio Caesare, consulibus Rubellio Gemino et Rufio [sic] Gemino mense Martio temporibus paschae, die octavo Kalendarum Aprilium, die primo azymorum quo agnum occideret ad vesperam, sicut a Moyse fuerat pruecepum [ed. H. Trankle]) and Hippolytus (ca. 170-ca. 236), Commentary on Daniel IV.23.3 ([upsilon][pi][alpha][tau][epsilon]upsilon]ov[tau]o[zita] 'Po[upsilon][phi]o[upsilon] [kappa][alpha]i 'Po[upsilon][beta][epsilon][lambda][lambda]i[omega]vo[zeta] [ed. G. N. Bonwetsch; also M. Lefevre]). Some manuscripts of the latter inexplicably append the consuls of 41 c.e. to those of 29 C.E. Among the many others are Lactantius (ca. 240-ca. 320), Divine Institutions IV.10 and On the Death of the Persecuted 2.1, and Augustine (354-430), City of God XVIII.5. Ideler's (1825-26, 2: 413--22) survey of classical sources on the topic at hand is still adequate. According to Ideler, the most thorough account up to his time is abbot Sanclement e's treatise entitled Exercitatio chronologica de anno domenicae passionis, appended to his four-volume De vulgaris aerae emendatione (Rome 1793). I have not been able to gain access to a copy. Both Sanclemenre and Ideler give the consul date pride of place and therefore favor 29 C.E. as the year of the event.

(2.) If SE 1 (cuneiform) begins in spring 311 B.C.E. (-310), then SE 340 (cuneiform) (1 + 339) begins in 29 C.E. (+29), that is, -310 + 339. SE 340 (Syriac) then begins in fall 28 C.E.

(3.) A recent guide with clear, precise, and satisfying detail on Jerwish ritual practice as it centers around the Talmud is Adin Steinsaltz's reference guide to his edition of the Talmud (Steinslaz 1989; for [LANGUAGE NOT REPRODUCIBLE IN ASCII] Pesakh "Passover," see p. 246). The time of the sacrifice is sometimes described as [LANGUAGE NOT REPRODUCIBLE IN ASCII] (Exodus 16:12; Leviticus 23:5), which literally means something like "between the two evenings." Mahler (1916: 36) cites three interpretations of this term: (1) time before and after sunset; (2) ninth to eleventh hour (ca. 3:00-5:00 p.m.) (Pharisees followed by modern Jewish custom); (3) time between sunset and complete darkness (Samaritans). Mahler prefers the third alternative.

(4.) One may check whether these were Fridays by means of their j(ulian) d(ay) numbers. The j.d. count is a continuous day count widely used in astronomy and beginning conventionally with 1 January 4713 B.C.E. a day that lasts from j.d. 0.00 (midnight between 31 December and 1 January) to j.d. 1.00 (mid-night between 1 January and 2 January); noon of that day is j.d. 0.50. Because that day is a Monday, j.d. numbers that produce a remainder of 0 when divided by 7 are Mondays. When the remainder is 4, they are Fridays. 18 March 29 C.E. is j.d. 1 731 727 and 15 April 28 C.E. is j.d. 1 731 755 (Schram 1908: 34). Both of these j.d. numbers produce a remainder of 4 when divided by 7.

(5.) In the two most extensive treatises on chronology surviving from antiquity, the natural day is defined as tempus ab oriente sole ad solis occasum "the time from sunrise to sunset" (Censorinus, De die natali XXIII) and [LANGUAGE NOT REPRODUCIBLE IN ASCII] "the time from sunrise to sunset" (Geminus, Introduction to the Phaenomena VI.1).

It was Joseph Scaliger who, in 1583, first proposed counting years with 4713 B.C.E. as year 1. Contrary to what is implied in almost all treatises on chronology, the day count is not his, but dates to the nineteenth century. The year count was created by a historian; the day count was made popular by astronomers.

(6.) For the Talmud on time-reckoning, see Zuckermann (1882).

(7.) What happens at the turn of the month is discussed in light of the calendrical practice of Yom Toy Sheni Shel Galuyyot in Depuydt 2002b.

(8.) For the history of this value, see Depuydt 2002a. The value is Babylonian in origin. It is part of the lunar theory called system B and was used by Greek astronomers, notably by Ptolemy in his Almagest (second century C.E.). Its sexagesimal notation is 29;31,50,8,20 days, that is, 29 + 31/[60.sup.1]+ 50/[60.sup.2] + 8/[60.sup.3] + 20/[60.sup.4]) days. The value differs by less than a second from the true mean value.

(9.) 27 April is j.d. 1 731 402, which divided by 7 gives a remainder of I (cf. note 4).

(10.) Hinz also speculates (1992: 56) that the very minute of Jesus's death was the exact time of opposition (the moment when the earth is right between the moon and the sun), namely 3.08 p.m., 27 April 28 C.E. (Goldstine 1973: 86). Is this a reference to the darkness in the land reported in the gospels (e.g., Matthew 27:45)? It is true that darkness can occur when sun, moon, and earth are on a single line. But that happens only at solar eclipses, which fall at new moon at the beginning of the Lunar month, when the moon is right between the sun and the earth, not at full moon in the middle of the lunar month, when the earth stands between sun and moon. Jesus died in the middle of the month. Eusebius already attributed the darkness to an eclipse, overlooking the fact that only solar eclipses can cause such an event (Turner 1910: 890-91),


Aime-Giron, N. 1931. Textes arameens d'Egypt. Cairo.

Amadon, Grace. 1942. "Ancient Jewish Calendation." Journal of Biblical Literature 61: 227-80.

Beaulieu, Paul-Alain. 1993. "The Impact of Month-lengths on the Neo-Bayblonian Cultic Calendar." Zeitschrift fur Assyriologie und Vorderasiastische Archaologie 83: 66-87.

Bickerman, Elias J. 1980. Chronology of the Ancient World, 2nd edn. Ithaca: Cornell Univ. Press.

Bilfinger, Gustav. 1888. Der burgerliche Tag: Untersuchungen uber den Beginn des Kalendertages im Classischen Alterium und im Christlichen Mittelalter. Stuttgart: Kohlhammer.

Bokser, Baruch M. 1987. "Was the Last Supper a Passover Seder?" Bible Review 3.2: 24-33.

Brack-Berasen, Lis. 1997. Zur Enstehung der babylonischen Mondtheorie: Beobachtung und theoretische Berechnung Von Mondphasen. Boethius, 40. Stuttgart: Franz Steiner Verlag.

Chwolson, Daniel. 1908. Dos letzte Passamahl Christi und der Tag seines Todes, nach den in Ubereinstimmung gebrachten Berichten der Synoptiker und des Evangelium Johannis, 2nd edn. Leipzig: H. Haessel.

Cowley, Arthur Ernest. 1923. Aramaic Papyri of the 5th Century B.C. Oxford: Clarendon Press.

Depuydt, Leo. 1996. "The Egyptian and Athenian Dates of Meton's Observation of the Summer Solstice." Ancient Society 27: 27-45.

-----. 1997. Civil and Lunar Calendar in Ancient Egypt. Orientalia Lovaniensia Analecta, 77. Leuven: Uitgeverij Peeters and Departement Oosterse Studies.

-----. 2002a. "History of the heleq." Forthcoming in the acts of the conference "Under One Sky" on ancient astronomy and mathematics held at the British Museum in June 2001, AOAT.

-----. 2002b. "Ancient and Medieval Sources and Mechanism of the Calendrical Practice of Yom Toy Sheni Shel Galuyyot." In Daily Life in the Ancient Near East. Bethesda, Md.: CDL Press.

Epping, Joseph. 1890. "Die babylonische Rechnung des Neumondes." Stimmen aus Maria-Laach 39: 225-40.

Fotheringham, J. K. 1908. "Calendar Dates in the Aramaic Papyri from Assuan." Monthly Notices of the Royal Astronomical Society 69: 12-20, with errata at 470.

-----. 1911. "A Reply to Professor Ginzel on the Calendar Dates in the Elephantine Papyri." Monthly Notices of the Royal Astronomical Society 71: 661-63.

-----. 1934. "The Evidence of Astronomy and Technical Chronology for the Date of the Crucifixion." Journal of Theological Studies 35: 146-62.

Gerhardt, O. 1930. "Das Datum der Kreuzigung Jesu Christi." Astronomische Nachrichten 240: 137-62.

Ginzel, Friedrich Karl. 1906-14. Handbuch der mathematischen und technischen Chronologie: Das Zeitrechnungswesen der Volker. 3 vols. (1906, 1911, 1914). Leipzig: J. C. Hinrichs'sche Buchhandlung.

Goldstein, Morris. 1950. Jesus in the Jewish Tradition. New York: Macmillan.

Goldstine, Herman H. 1973. New and Full Moons 1001 B.C. to A.D. 1651. Memoirs of the American Philosophical Society, 94. Philadelphia: American Philosophical Society.

Hinz, Walther. 1989. "Chronologie des Lebens Jesu." Zeitschrift der Deutschen Morgenlandischen Gesellschaft 139: 301-9.

-----. 1992. "Jesu Sterbedatum." Zeitschrift der Deutschen Morgenlandischen Gessellschaft 142: 53-56.

Horn, S. L., and L. H. Wood. 1954. "The Fifth-century Jewish Calendar at Elephantine." Journal of Near Eastern Studies 13: 3-20.

Ideler, Ludwig. 1825-26. Handbuch der mathematischen und technischen Chronologie, aus den Quellen bearbeitet. Berlin: August Rocker.

Klausner, Joseph. 1925. Jesus of Nazareth: His Life, Times, and Teaching. New York: Macmillan.

Koestermann, Erich. 1965. Cornelius Tacitus: Annalen, Band II, Buch 4-6. Heidelberg: Carl Winter.

Kraeling, Emil Gottlieb Heinrich. 1953. The Brooklyn Museum Aramaic Papyri: New Documents of the Fifth Century B.C. from the Jewish Colony at Elephantine. New Haven: Yale Univ. Press.

Kugler, Franz Xaver. 1907-24. Sternkunde und Sterndienst in Babel. 2 vols. (1 [1907], 2.1 [1909/1910], 2.2.1 [1912]. 2.2.2 [1924]). Munster: Aschendorffsche Verlagsbuchhandlung. Two supplements appeared in 1913 and 1914; a third, by Johann Schaumberger, in 1935.

Loi, V. 1971. "Il 25 marzo data pasquale e la cronologia giovannea della Passione in eta patristica." Ephemerides Liturgicae 85: 48-69.

Mahler, Eduard. 1916. Handbuch der judischen Chronologie. Grundriss (ler Gesamtwissenschaft des Judentums. Leipzig: Gustav Fock.

Nilsson, Martin P. 1920. Primitive Time-reckoning: A Study in the Origins and First Development of the Art of counting Time among the Primitive and Early Culture Peoples. Lund: C.W.K. Gleerup.

Olmstead, Albert Ten Eyck. 1942. Jesus in the Light of History. New York: Scribner's.

Parker, Richard A. 1955. "Some Considerations on the Nature of the Fifth-century Jewish Calendar at Elephantine." Journal of Near Eastern Studies 14: 271-74.

Parker, Richard A., and Waldo H. Dubberstein. 1956. Babylonian Chronology. Brown University Studies, 19. Providence: Brown Univ. Press.

Pepin, M. Barlow. 1996. "In Quest of the Youngest Moon." Sky and Telescope December 1996: 104-6.

Porten, Bezalel. 1990. "The Calendar of Aramaic Texts from Achaemenid and Ptolemaic Egypt." In Irano-Judaica, II: Studies Relating to Jewish Contacts with Persian Culture throughout the Ages. Ed. Shaul Shaked and Ammon Netzer. Pp. 13-32. Jerusalem: Makhon Ben-Tsevi.

Porten, Bezalel, and Ada Yardeni. 1986-93. Textbook of Aramaic Documents from Ancient Egypt. 3 vols. (1986, 1989, and 1993). I: Letters (Appendix: Aramaic Letters from the Bible). II: contracts. III: Literature, Accounts, Lists. Jerusalem: Akademon.

Pritchett, W. Kendrick. 1982. "The Calendar of the Gibbous Month." Zeitschrift fur Papyrologie und Epigraphik 49: 243-66.

Sachs, Abraham J., and Hermann Hunger. 1988. Astronomical Diaries and Related Texts from Babylonia, I: Diaries from 6S2 B.C. to 262 B.C. Osterreichische Akademie der Wissenschaften, Philosophisch-historische Klasse, 195. Vienna: Verlag der Akademie.

Schaefer, Bradley E.; Imad A. Ahmad; and Leroy E. Doggett. 1993. "Records for Young Moon Sightings." Quarterly Journal of the Royal Astronomical Society 34: 53-56.

Schram, Robert and Gabriele. 1908. Kalendariographische und Chronologische Tafeln. Leipzig: J. C. Hinrichs'sche Buchhandlung.

Schwarz, Adolf. 1872. Der judische Kalender historisch und astronomisch untersucht. Breslau: Schletter'sche Buchhandlung.

Steinsaltz, Adin. 1989. The Talmud: The Steinsaltz Edition; A Reference Guide. New York: Random House.

Strauli, Robert. 1991. "Herkunft und Bedeutung unseres Kalenders." In Museion 2000. Zurich: ABZ Verlag.

Turner, Cuthbert Hamilton. 1910. "Bible." In Encyclopaedia Britannica, 11th ed., vol. 3, Pp. 849-94. Cambridge: Cambridge Univ. Press.

Wacholder. Ben Zion, and David B. Weisberg. 1976. "Visibility of the New Moon in Cuneiform and Rabbinic Sources." In B. Z. Wacholder, Essays on Jewish Chronology and Chronography. Pp. 59-74. New York: Ktav Publishing House.

Wilcken, Ulrich. 1912. Grundzilge und Chrestomathie der Papyruskunde. vol. 1.2. Leipzig and Berlin: Teubner.

Zuckermann, Benedict. 1882. Materialien zur Eutwickelung der altjudischen Zeitrechnung im Talmud. Jahres-Bericht des judisch-theologischen Seminars "Fraenkel'scher Stiftung." Breslau: S. Schottlaender.
COPYRIGHT 2002 American Oriental Society
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2002 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Depuydt, Leo
Publication:The Journal of the American Oriental Society
Date:Jul 1, 2002
Previous Article:The high places (bamot) and the reforms of Hezekiah and Josiah: an archaeological investigation.
Next Article:IBN al-Salah al-Shahrazuri and the isnad.

Terms of use | Copyright © 2017 Farlex, Inc. | Feedback | For webmasters