Science in Christian Perspective




The Date of the Crucifixion


Department of Metallurgy and Science of Materials
University of Oxford
Parks Road, Oxford, England

From: JASA 37 (March 1985): 2-10.

*Also Jesus College, Oxford. * *Also Department of Astrophysics, University of Oxford.

The date of the crucifixion has been debated for many years yet there has been no agreement on the year nor the day on which Jesus died. In this review astronomical calculations are used to reconstruct the first century A.D . Jewish calendar and to date a lunar eclipse which Biblical and other references suggest followed the crucifixion. The evidence strongly points to Friday, 3 April, A.D. 33 as being the date when Christ died.

Chronology is the backbone of history. The key date in the chronology of the life of Jesus is the date of the crucifixion, since it is directly relevant to the length of his ministry, and to the date and nature of the Last Supper. We believe that the only way to establish the precise date of the crucifixion is by using science, history and theology in an interdisciplinary study. Previous attempts to date the crucifixion have used a process of elimination, showing that every year other than the one chosen is incompatible with the available evidence (see, for example, Hoehner,1 Finegan2). In this article we consider the first positive dating of the crucifixion using lunar eclipse evidence."' We also reassess the 'elimination method' using a new technique for reconstructing the first century A.D. Jewish calendar, which should be more accurate than previous versions. Details of our work have already been published.3,4 In this paper we bring together these details and add some further comments.

Biblical Background

There are three main pieces of evidence for dating the crucifixion:

(i) Jesus was crucified when Pontius Pilate was procurator of Judaea (all four Gospels; also Tacitus5),2
which is well documented to be A.D. 26-36.

(ii) All four Gospels agree that Jesus died a few hours before the commencement of the Jewish Sabbath, i.e. he died before nightfall on a Friday.

(iii) All four Gospels agree to within about a day (see below) that the crucifixion was at the time of Passover.

These three pieces of evidence compel us to reject many of the dates which have been suggested in the past for the crucifixion. For example, one of the earliest traditions, going back to Tertullian (A.D. 200), gives the date as A.D. 29, March 25. This date was not accepted everywhere throughout the early church and we now know from astronomical calculations that the time of the Passover moon in A.D. 29 was in April, not March.

In the official festival calendar of Judaea, as used by the priests of the temple, Passover time was specified precisely (see, for example, Reicke6). The slaughtering of the lambs for Passover occurred between 3pm and 5pm on the 14th day of the Jewish month Nisan (corresponding to March/April in our calendar). The Passover meal commenced at moonrise that evening, i.e., at the start of 15 Nisan (the Jewish day running from evening to evening) (Leviticus 23 v. 5; Numbers 28 v. 16). There is an apparent discrepancy of one day in the Gospel accounts of the crucifixion which has been the subject of considerable debate. In John's Gospel, it is stated that the day of Jesus' trial and execution was the day before Passover (John 18 v. 28 and 19 v. 14), Hence John places the crucifixion on 14 Nisan. The correct interpretation of the Synoptics is less clear and we consider briefly three of the many possible interpretations which have been proposed.

(a) A straightforward reading of the Synoptics would seem to indicate that the Last Supper was a Passover meal, eaten at Passover time (i.e. in the evening at the start of 15 Nisan) with the crucifixion occurring later that Jewish day, i.e on 15 Nisan (e.g. Mark 14 v. 12). This disagrees with John's date of 14 Nisan (see Jeremias7).

(b) Many scholars propose that the Last Supper described by the Synoptics was not a strict Passover meal. It is suggested that Jesus, knowing of his imminent arrest, held a Passover-like meal on the evening before Passover (see Luke 22 v. 15). Supporters of this interpretation note that the Synoptics make no mention of a Passover lamb being slain and roasted for the Last Supper. This interpretation is in broad agreement with the Johannine account in which the farewell meal is explicitly stated to have occurred before the feast of Passover (John 13 v. 1). The timing also agrees, so that on this theory all 4 gospels give 14 Nisan as the crucifixion date. A number of variations on this basic interpretation exist (e.g. references 7,8,9).

(c) Jaubert10 has proposed that the Last Supper reported by the Synoptics was a strict Passover Meal but held at Passover time as calculated using the 'sectarian' calendar of the Qumran community and others. According to this theory the Last Supper was held on Tuesday evening, i.e. at the start of the Jewish Wednesday (the sectarian calendar Passover day, and recorded by the Synoptics) the crucifixion was on Friday (all 4 Gospels) and the official Passover was on Saturday (recorded by John). (For a discussion of calendars in use in the first century A.D. see for example, refs. 2 and 11). According to this theory all four Gospels again give 14 Nisan (official calendar) as the crucifixion date.

Thus some scholars believe that all 4 Gospels place the crucifixion on Friday, 14 Nisan, others believe that according to the Synoptics it occurred on Friday, 15 Nisan. For generality at this stage we assume that both dates may be possible. The problem that then has to be solved is that of determining in which of the years A.D. 26-36 the 14th and 15th Nisan fell on a Friday. As is well known, various authors (e.g. 7, 12, 13, 14, 15) have attempted to use astronomy to provide a solution to this problem. This is not entirely straightforward however, since, although astronomical calculations can accurately specify the times of new and full moons we do

Thus there is an impressive unanimity from all sources that the crucifixion was on 14 Nisan and consequently the only two plausible years for the crucifixion are A.D. 30 and A.D. 33.

not know with what skill the Jews of the first century could detect the first faintly glowing lunar crescent following conjunction with the sun (the new moon itself being invisible, of course).

Reconstructing the First Century A.D.
Jewish Calendar

In the past it appears to have been common practice to assume arbitrarily that the sickle of the new moon would be invisible to the unaided eye until a certain length of time (usually 30 hours) had elapsed since conjunction. Fotheringham14 applied a more realistic criterion, based on the apparent position of the moon in the sky at sunset, to the problem of the visibility of small lunar crescents. Maunder15 modified and improved Fotberingham's criterion. Even Maunder's limit of visibility is not rigorous as several thin crescents have been observed which would have been deemed impossible using his criterion.

In the present work we have computed the lunar crescent visibility as a function of time after sunset for the beginning of each lunar month in the period of interest. In order to do this, the lunar semi-diameter and the position of the moon in the sky at and after sunset are first evaluated from harmonic synthesis of the perturbed orbits of the earth and moon. The sky brightness for an observer at Jerusalem is then calculated as a function of the depression of the sun below the horizon, as is the moon's apparent surface brightness. Whether or not the moon is visible depends upon whether its contrast with the sky background exceeds the visual contrast threshold.16 For the latitude of Jerusalem, our criteria for the first visibility of the lunar crescent corresponds to the lunar altitude at sunset being approximately 0.50 lower than that given by Maunder's criterion. This approach gives results which

are consistent with many recent observations of the first sickle of the new moon and we can therefore have considerable confidence in the calculations. Using this method, and assuming normal atmospheric transparency, we obtain Table 1, in which all cases close to the visibility limit are noted in the footnotes.

Although in the first century A.D. the beginning of the Jewish lunar month (in the official calendar) was fixed rigorously by astronomical observation, there remains an uncertain calender factor: that of intercalary (or leap) months. Twelve lunar months total approximately 11 days less than a solar year. For agricultural and religious festival reasons the Jews kept lunar months at approximately the same place in the solar year by the intercalation of a thirteenth month when necessary (one intercalary month being required approximately every three years).

Different methods of intercalation were used at different periods of Jewish history but in the first century A.D. intercalation was regulated annually by proclamation by the Sanhedrin according to certain criteria (7, 11, 12, 17). The most important of these was that Passover must fall after the vernal equinox. [It is clear from the Talmudic tract Sanhedrin 10b-13b that the Jews knew in advance when the equinox would be. We do not know the method they employed for this. There are various possibilities, for example use of the so-called 'gates of heaven' (Neugebauer36), or use of the time of the heliacal risings of particular groups of stars (e.g. the Egyptian decans) and the position of the new crescent relative to these asterisms (especially the Pleiades, as in Babylonia) to determine in advance when the equinox would be]. If it was noticed towards the end of a Jewish year that Passover would fall before the equinox the intercalation of an extra month before Nisan was decreed. Table 1 has been constructed on this basis and therefore includes intercalary months. However, a leap month could be decreed if the crops had been delayed by unusually bad weather (since the first fruits must be ripe for presentation on Nisan 16) and if the lambs were too young. Unfortunately we possess no historical reports as to the proclamation of leap-months in the years A.D. 26-36. it is therefore possible that in some years Nisan was one month later than given in Table 1, on account of unusually severe weather. Calculations show that in the period A.D. 26-36, if Nisan was one month later than given in Table 1, Nisan 14 would not fall on a Friday in any year and Nisan 15 would only fall on a Friday in A.D. 34 (April 23).

Possible Dates for the Crucifixion

2 lists the possible dates of a Friday crucifixion, on 14 or 15 Nisan, obtained from Table 1 and including April 23, A.D. 34. These dates in Table 2 are the only

Colin J. Humphreys received a B.Sc. (Physics) from London University and a Ph.D. (Metal Physics) from Cambridge University. He is currently the Chairman of the Metallurgy and Science of Materials Faculty at Oxford University and a Fellow of Jesus College, Oxford. He is also Chairman of the Commission on Electron Diffraction of the International Union of Crystallography. His primary research interests are electron diffraction and microscopy, semiconductor materials and electron beam lithography. He has a personal interest in science and Christian belief.

W. Graeme Waddington received a B.Sc. (Physics) from Durham University and a D.Phil (Astrophysics) from Oxford University. His primary research interest is Solar Physics. He is working with Colin Humphreys on computing electron microscope images.

ones that are astronomically and calendrically possible for the crucifixion. We now examine these dates in the light of other evidence.

A.D. 27 is almost certainly too early for the crucifixion. Luke 3:1-2 carefully states that John the Baptist commenced his ministry in the fifteenth year of Tiberius Caesar (Jesus was baptised by John subsequently). Depending on whether Hellenistic (Roman) civil or the Jewish ecclesiastical reckoning is used, the fifteenth year (=340 Seleucid Era) is autumn A.D. 28-29 or spring A.D. 29-30 (Edwards,18 also private communication). In addition, most scholars believe that Pilate had been procurator for some time before the crucifixion (see Luke 13:1 and 23:12). These two points rule out the possibility of an A.D. 27 crucifixion.

A.D. 34 is almost certainly too late for the crucifixion since it would conflict with the probable date of Paul's conversion (A.D. 34, see ref. 19). In addition, A.D. 34 is only a possibility if the weather was exceptionally severe. There is no positive evidence in favour of A.D. 34 and we therefore rule it out. (The only eminent supporter of 23 April, A.D. 34 that we have come across was Sir Isaac Newton, and his main reason seems to have been that 23 April is St George's Day!).

Having eliminated A.D. 27 and A.D. 34 as possible years for the crucifixion, we note from Table 2 that the crucifixion must have occurred on 14 Nisan, and that the previously listed interpretation (a) of the Last Supper cannot be correct. It is perhaps also worth noting that science has been used to distinguish between different theological interpretations of the nature of the Last Supper, and has shown on calendrical grounds that the Last Supper cannot have been a Passover meal held at the official time. In addition, we have shown that the crucifixion occurred on 14 not 15 Nisan. Thus Jesus died at the same time as the Passover lambs were slain. This is consistent with many New Testament statements, for example, 'Christ our Passover is sacrificed for us' (1 Cor. 5:7). In addition, Paul refers to Christ as the first fruits of those who rise from the dead (1 Cor. 15:20), a clear analogy with the offering of the first fruits in the temple, which occurred on Nisan 16. Paul would surely not have used this analogy had the crucifixion been on Nisan 15 and the resurrection on Nisan 17. Thus in describing Christ symbolically as the Passover lamb and as the first fruits, the Pauline chronology of the crucifixion events is identical to that of John. Both are consistent with the Synoptic chronology, provided the Last Supper was not a Passover meal held at the official time. In addition, the Babylonian Talmud records that Jesus' death was on the eve of Passover, i.e. on 14 Nisan (Sanhedrin 43a). Thus there is an impressive unanimity from all sources that the crucifixion was on 14 Nisan and consequently the only two plausible years for the crucifixion are A.D. 30 and A.D. 33.

The earliest Possible date for the commencement of the ministry of Jesus is autumn A.D. 28 (see Edwards18) and John's gospel records three different Passovers; occurring in the ministry (including the one at the crucifixion). Hence, if this evidence is accepted, A.D. 30 cannot be the crucifixion year, leaving A.D. 33 as the
only possibility. A.D. 33 is also consistent with the I temple reference.' At the first Passover of Jesus' ministry John 2:20 records that the Jews said to Jesus 'It has taken 46 years to build this temple.' Assuming this refers to the inner temple (see Hoehner1), the 46 years leads to A.D. 30 or 31, depending upon how much preparation time was involved before building commenced. If the only Passovers of Jesus' ministry were

We conclude that the words of Cyril and the Report of Pilate may be used as secondary evidence supporting our interpretation of the words of Peter that the moon appeared like blood on the evening of the crucifixion.

the 3 explicitly mentioned in John's gospel, an A.D. 33 crucifixion implies a ministry of about 21/2 years. Many scholars believe that John omitted to mention a further Passover, so that the ministry was for 31/2 years.

This date, 3 April A.D. 33, is supported by many scholars (e.g. Hoehner,1 Reicke6, Ogg17). However, not all scholars accept that A.D. 33 is preferable to A.D. 30, and the date 7 April A.D. 30 is also strongly supported (e.g. Finegan 2 Bruce,' Robinson 21). Without further evidence it does not seem possible to decide conclusively between these two dates, although 3 April A.D. 33 is considered to be much the more probable for the reasons given above. A few scholars support dates other than A.D. 30 or A.D. 33 although these do not seem calendrically possible. We now consider the new evidence presented in our previous papers,3,4 which provides the first positive dating of the crucifixion.

The Moon Turned to Blood

The new evidence concerns the meaning and signif icance of the moon being 'turned to blood,' referred to in the Bible and elsewhere. In Acts Chapter
2 v. 14-21 it is recorded that on the day of Pentecost the apostles were accused by a crowd of being drunk. Peter stood up and said 'No, this is what was spoken by the prophet Joel: In the last days, God says, I will pour out my spirit on all people ... I will show wonders in the heavens above ... The sun will be turned to darkness and the moon to blood before that great and glorious day of the Lord shall come. And everyone who calls on the name of the Lord shall be saved.'

Commentators are divided upon whether Peter was claiming that all the quoted prophecy from Joel had recently been fulfilled (e.g. Neil21) or whether the words refer to the future. We will investigate the former interpretation further, demonstrate that 'the moon turned to blood' probably refers to a lunar eclipse, and show that this interpretation is selfconsistent and enables the crucifixion to be dated precisely.

Peter prefaces his quotation from Joel with the words 'Let me explain this to you ... this is what was spoken by the prophet Joel.' Peter therefore appears to be arguing that recent events have fulfilled the prophecy he is about to quote. If this interpretation is correct 'the last days' (v. 17) began with Christ's first advent (for similar usage see 1 Peter 1:20; Hebrews 1:1-2) and the outpouring of the spirit (v. 17-18) commenced at Pentecost; 'that great and glorious day' (v. 20) refers to the resurrection, since which time 'everyone who calls on the name of the Lord will be saved' (v. 2 1). 'The sun will be turned to darkness' (v. 20) refers back to the 3 hours of darkness which occurred only 7 weeks previously, at the crucifixion (Matthew 27:45), and would be understood as such by Peter's audience. As is well known, the mechanism by which the sun was darkened may have been a khamsin dust storm (see refs. 22 and 23). Since the darkened sun occurred at the crucifixion it is reasonable to suppose that 'the moon turned to blood' occurred that same evening, 'before that great and glorious day,' the resurrection.

The interpretation of Acts 2:20 suggested above, that 'The sun will be turned to darkness and the moon to blood'refers back to the crucifixion, is supported by the New Testament scholar, F. F. Bruce,24 who states in his commentary on the Acts 'Peter's hearers may have associated the phenomena in (Acts 2) vv. 19 f. with those which attended the preternatural darkness on Good Friday.'

There is some other evidence that on the day of the crucifixion the moon appeared like blood. The so-called Report of Pilate, a New Testament Apocryphal fragment (see James25), states 'Jesus was delivered to him by Herod, Archelaus, Philip, Annas, Caiphas, and all the people. At his crucifixion the sun was darkened; the stars appeared and in all the world people lighted lamps from the sixth hour till evening; the moon appeared like blood.'

Much of the New Testament Apocrypha consists of highly theatrical literature, which cannot be used as primary historical evidence. Tertullian records that Pilate wrote a Report of all the events surrounding the crucifixion and sent this to the Emperor Tiberias. The manuscript fragments that we possess of the Report of Pilate are all of later date but may be partly based on this very early lost document (see James'). If this is the case the Report may provide independent evidence that 'the moon appeared like blood' following the crucifixion. On the other hand, the Report may have used the Acts as a source and not be independent from it. If this is the case, however, the event described by Peter 'the moon turned to blood' is clearly stated in the

There is strong evidence that when Peter and the Report of Pilate refer to the moon turning to blood on the evening of the crucifixion they are describing a lunar eclipse.


Report to have occurred at the crucifixion. A third possibility, which is the most likely, is that the so-called Report is a late Christian 'forgery.' If this is correct, there must have been a Christian tradition that at the crucifixion the moon appeared like blood.

Further evidence is provided by Cyril of Alexandria, the Patriarch of Alexandria in A.D. 412. After stating that there was darkness at the crucifixion he adds I something unusual occurred about the circular rotation of the moon so that it even seemed to be turned into blood.' He notes that the prophet Joel foretold this sign." We conclude that the words of Cyril and the Report of Pilate may be used as secondary evidence supporting our interpretation of the words of Peter that the moon appeared like blood on the evening of the crucifixion.

A Lunar Eclipse Following the Crucifixion

'The moon turned to blood' is a graphic description of a lunar eclipse. The reason an eclipsed moon is blood red is well known and the effect has been well documented. Even though during an eclipse the moon is geometrically in the earth's shadow, some sunlight still reaches it by the refraction of light passing through the earth's atmosphere. This light is red since it has traversed a long path through the atmosphere and scattering by air molecules and very small particles preferentially removes the blue end of the spectrum. The combination of scattering and refraction produces the deep blood-red colour of a lunar eclipse.

'The moon turned to blood' has been used by writers and historians to describe lunar eclipses for many centuries, and the expression dates back to at least 300 B.C. Descriptions of some well documented ancient eclipses have been compiled by Ginzel' and matched with calculated eclipse dates. We quote three examples:
                                                                                                                                                                                                                                                                                             (i)The lunar eclipse of 20 September, 331 B.C. occurred two days after Alexander crossed the Tigris and the moon was described by Curtius (IV, 10 (39), 1) as 'suffused with the colour of blood.'

(ii) The lunar eclipse of 31 August A.D. 304 (probably) which occurred at the martyrdom of Bishop Felix, was described in Acta Sanctorem. 'when he was about to be martyred the moon was turned to blood.'

(iii) The lunar eclipse of 2 March A. D. 462 was described in the Hydatius Lemicus Chronicon thus I on March 2 with the crowing of cocks after the setting of the sun the full moon was turned to blood.'

In the mediaeval European annals compiled by Pertz 28 there are so many lunar eclipses described by I the moon turned to blood' that the phrase appears to be used as a standard description. Stephenson' considers that the prophecy of Joel (2 v. 31) of the moon turning into blood (which Peter quoted at Pentecost) clearly alludes to a lunar eclipse.

There is therefore strong evidence that when Peter and the Report of Pilate refer to the moon turning to blood on the evening of the crucifixion they are describing a lunar eclipse. It is surprising that this deduction does not appear to have been made before, although Bruce36 almost reaches this conclusion. He states, with reference to Peter's Pentecost speech, 'it was little more than seven weeks since the people in Jerusalem had indeed seen the sun turned to darkness, during the early afternoon of the day of our Lord's crucifixion. And on the same afternoon the paschal full moon may well have appeared blood-red in the sky in consequence of that preternatural gloom.' Presumably Bruce and other commentators have not been aware that a blood-red moon is a well-documented description of a lunar eclipse.

Lunar Eclipses Visible from Jerusalem A.D.

We have determined the eclipses relevant to our work using the most comprehensive data availableP' as corrected by Stephenson (private communication, see also refs. 32 and 33), who used ancient Babylonian eclipse records to refine the calculations to take into account more accurately long term changes in the earth's rate of rotation. These calculations agree very well with records of Babylonian and Chinese eclipse observations (the probable error in a calculated eclipse time 2000 years ago is typically 5 minutes). All lunar eclipses (total and partial) visible from Jerusalem between A.D. 26-36 are listed in Table 3.

From Table 3, in the period A.D. 26-36 there was one, and only one, lunar eclipse at Passover time visible from Jerusalem, namely that of Friday, 3 April, A.D. 33. This date is the most probable date for the crucifixion deduced independently using other data. The interpretation of Peter's words in terms of a lunar eclipse is therefore not only astronomically and calendrically possible, but it also allows us with reasonable certainty to specify Friday, 3 April, A.D. 33 as being the date of the crucifixion. The random probability of a lunar eclipse occurring at moonrise (see below) on a particular date is, of course, small.

It is interesting to note that there have been a few references to this eclipse in the past, for example Hind" calculated that there was a lunar eclipse on 3 April A.D. 33, however his calculations showed that this eclipse was not visible from Jerusalem and presumably it was considered irrelevant to the date of the crucifixion. It is only recently that we have been able to take into account accurately the effects of long term changes in the earth's rate of rotation. Hence it is only in recent years that it has become possible to state with confidence that this lunar eclipse was visible from Jerusalem: that is its importance.

The Lunar Eclipse on Friday, 3 April, A.D. 33

Calculations show that this eclipse was visible from Jerusalem at moonrise. All times quoted below are local Jerusalem times as measured by a sundial. The start of the eclipse was invisible from Jerusalem, being belowthe horizon. The eclipse began at 3.40pm and reached a maximum at about 5.15pm, with 60% of the moon eclipsed. This was also below the horizon from Jerusalem. The moon rose above the horizon, and was first visible from Jerusalem, at about 6.20pm (the start of the Jewish Sabbath and also the start of Passover day in A.D. 33) with about 20% of its disc in the umbra of the earth's shadow and the remainder in the penumbra. The eclipse finished some 30 minutes later at 6.50pm.

Although at moonrise only 20% of the total area (irr') of the moon's disc was eclipsed (i.e. in the umbral shadow), calculations show that this 20% 'bite' was positioned close to the top (i.e. leading edge) of the moon. Fig. 1 shows the appearance of the moon at, and shortly after, moonrise on 3 April A.D. 33. As the umbral shadow (in which the sun is geometrically entirely hidden) was near the top of the moon, about two-thirds of the visible area of the rising moon would initially be seen as eclipsed (see bottom of Fig. 1), while the remainder would have been in the penumbral shadow.

There is great variability in exact colouration from eclipse to eclipse owing to atmospheric conditions. As explained previously, the umbral shadow is normally blood red. However, this colour is most noticeable during total lunar eclipses. For partial eclipses, particularly with the moon at high altitude, there is a large contrast difference between the obscured and unobscured part of the moon, so that the moon often appears almost white with a very dark 'bite' removed. However for some partial eclipses the red colour of the umbral

It is only recently that we have been able to take into account accurately the effects of long term changes in the earth's rate of rotation. Hence it is only in recent years that it has become possible to state with confidence that this lunar eclipse was visible from Jerusalem: that is its importance.

shadow is clearly visible. For example, Davis35 has recently depicted in colour an eclipse sequence as seen by the human eye with the moon low in the sky, and the colouration of the umbra in the partial eclipse phase is almost as vivid as when the eclipse is total.

For the case of the eclipse of 3 April A.D. 33 the moon was just above the horizon. The most probable colour of the rising moon would be red in the umbra] shadow (shaded in Fig. 1) and yellow-orange elsewhere. At moonrise the initially small yellow-orange region would indicate that the moon had risen, but with most of its visible area 'turned to blood.' If in fact a massive dust storm was responsible for darkening the sun a few hours previously, dust still suspended in the atmosphere would tend to modify the above colours. The nature of this modification would depend upon the size distribution of the particles, but they would probably further darken and redden the moon.

The majority of lunar eclipses pass unnoticed, occurring when we are asleep or indoors. This eclipse however would probably have been seen by most of the population of Israel, since the Jews on Passover Day would be looking for both sunset and moonrise in order to commence their Passover meal. Instead of seeing the expected full Paschal moon rising they would have initially seen a moon with a red 'bite' removed (Fig. 1). The effect would be dramatic. The moon would grow to full in the next half-hour. The crowd on the day of Pentecost would undoubtably understand Peter's words, the moon turning to blood, as referring to this eclipse which they had seen.

A Crucifixion Solar Eclipse?

Finally, we must consider the well-known reference to a solar eclipse in some translations of Luke 23: 44-45. A typical translation is 'It was now about the sixth hour, and darkness came over the whole land until the ninth hour: the sun was eclipsed.' The three Greek words translated 'the sun was eclipsed' occur in only 5 early (but major) manuscripts of Luke. Other manuscripts do not contain these three words. Consequently some translations of Luke's gospel refer to this eclipse and others do not. A solar eclipse is of course astronomically impossible at Passover (full moon) time. In addition solar eclipses last for minutes not hours.

We consider there are two possible explanations for this apparent reference to a solar eclipse, One is that the original Luke text (which we do not possess) did not refer to a solar eclipse, but that a scribe copying this text, and knowing the oral tradition of an eclipse at the crucifixion, wrongly concluded this was a solar eclipse and amended the text accordingly. This could explain why these words are not present in all Luke manuscripts. We develop this argument further in ref. 4. The other possibility is that the Greek words translated 'the sun was eclipsed' were not intended to refer to an eclipse of the sun by the moon in the scientific sense, but simply refer to the sun not being visible. Strong support for this interpretation comes from a Messianic section of the Sibylline Oracles23 which states 'And straightway dust is carried from heaven to earth and all the brightness of the sun fails at midday from the heavens.' The Greek word translated 'fails' by Charles23 is identical to that in Luke 23:45 and usually translated I eclipsed.' It is clear that the sun darkening mechanism referred to in the Oracles, this section of which was probably written before A.D. 160, is a dust storm and not an eclipse. We conclude that the apparent reference to a solar eclipse in some manuscripts of Luke's Gospel presents no problems.


Establishing the precise date of any ancient event is well-known to be extremely difficult. It is probably not an exaggeration to state that only when ancient chronology is based upon calculable astronomical phenomena can we have certainty and precision. In this review, astronomy has been applied in two different ways to date the crucifixion.

First, we have reconstructed the first century A.D. Jewish calendar using detailed astronomical calculations. The main textual evidence for the time of the crucifixion has been reviewed and we have concluded that only 2 dates, 7 April, A.D. 30 and 3 April, A.D. 33, fit the main pieces of evidence for when Christ died. Other textual evidence, more difficult to interpret correctly, strongly favours Friday, 3 April, A.D. 33 as the date of the crucifixion.

Secondly, we have calculated that there was a lunar eclipse visible from Jerusalem at moonrise on the evening of Friday, 3 April A.D. 33. If this date was indeed the date of the crucifixion we would expect

The crowd on the day of Pentecost would undoubtably understand Peter's words, the moon turning to blood, as referring to this eclipse which they had seen.

some textual reference to a lunar eclipse at the crucifixion. This paper presents three such textual references, from the Acts of the Apostles, the writings of Cyril of Alexandria and the Report of Pilate. There are therefore powerful convergent arguments pointing to 3 April A.D. 33 as being the date of the crucifixion.

The authors are most grateful to Mr. J. G. Griffith for information on some New Testament manuscripts and for detailed discussions, and to Dr. F. R . Stephenson for supplying very "curate lunar eclipse data and some important astronomical references. We are also grateful to the following historians' scientists and theologians for their comments on a first draft of this paperDr. 0. R. Barclay, Dr. G. A. D. Briggs, Mr. 0. Edwards, Canon J. Fenton, Dr. P. E. Hodgson, Dr. G. Vermes, Dr. D. Wenham and Dr. D. E. H. Whiteley.


1Hoehner, H.W., Chronological Aspects of the Life of Christ. Zondervan: Grand Rapids (1977).
Finegan, J., Handbook of Biblical Chronology. Princeton University Press: Princeton (1964).
3'Humphreys, C.J., and Waddington, W.G. Nature, 306, 743 (1983).
4'Humphreys, C.J., and Waddington, W.G. in the Finegan Festschrift (ed. J. Vardaman, E.M. Yamauchi and B. Van Eldeven). (1984).
5Tacitus, Annals, XV, 44.
6Reicke, B., The New Testament Era. Translated from the German by D. E. Green. Black: London (1968).
7Jeremias, J., The Eucharistic Words of Jesus, translated by N. Perrin. SCM Press: London (1966).
Bruce, F.F., New Testament History. Pickering and Inglis; London and Glasgow (1982).
Whiteley, D.E. H., Aufstieg und Niedergang der Romischen Welt, 25, Part 11, de Gruyter: East Berlin and New York, in Press (1983).
Ojaubert, A., La Date de la Cene. Gabalda: Paris (1957).
11Schurer, E., Vermes, G., and Millar, F., The History of the Jewish People in the Age of Jesus Christ, Vol. 1. Edinburgh University Press: Edinburgh (1973).
Fotheringham, J. K., J. Theological Studies, 35, 146 (1934).
13Goldstine, H, H., New and Full Moons, 1000 B.C. to A.D. 1651. Fortress Press: Philadelphia (1973).
Fotheringham, J.K., Mon. Not. R. Astr. Soc. 70,527 (1910).
15Maunder, E. W., J. Brit. Asir. Assoc. 21, 355 (1911).
16Kornfeld, G.H., and Lawson, W.R., J. Opt. Soc. Am. 61,811 (1971).

G, The Chronology of the Public Ministry of Jesus. CUP: Cambridge (1940).

Edwards, 0., Palestine Exploration Quarterly, p. 29 (1982).
19ewett, R., Dating Paul's Life. SCM Press: London (1979).
20Robinson, J.A.T., Redating the New Testament. SCM Press: London (1976).
21Neil, W., The Acts of the Apostles. Eerdmans: Grand Rapids, Michigan (1973).
22Driver, G.R., J. Theological Studies XVI, 327 (1965).

'Sibylline Oracles, 111, 800. In Charles, R. H., The Apocrypha and Pseudepigrapha of the Old Testament. Oxford University Press (1913).
24Bruce, F.F., The Acts of the Apostles. inter-Varsity Press: Leicester (1952).
25James, M.R., The Apocryphal New Testament. Clarendon Press: Oxford (1953), p. 154.
Pusey, P.E., Cyrilli archiop. Alex. in XII Prophetas, i, p. 341, in Joel, ii, 30-31.
Ginzel, F.K., Spezieller Kanon der Sonnen- and Mondfinsternisse. Mayer and Muller: Berlin (1899).
Pertz, G.H., Monumenta Germaniae Historica, Scriptores, Hanover (1859). Reprinted Kraus Reprint Corporation: New York (1963), Vols 16-19.
Stephenson, F.R., Veins Testamentum, 19, Z24 (1969).
Bruce, FF., Commentary on the Book of the Acts. Eerdmans: Grand Rapids, Michigan (1981).
Oppolzer, T.R. von, Kanon der Finsternisse (1887). Translated by 0. Gingerich. Dover: New York (1961).
Stephenson, F.R., and Clark, D.H., Applications of Early Astronomical Records. OUP: Oxford (1978).
Stephenson, F.R., Scientific American (1982), p. 154.
34Hind, J.R., Nature, 6, 251 (1872).
35Davis, D., Sky and Telescope, 64, 391 (1982).
36Neugebauer, 0., Orientalia, 33, 49 (1964).