Lunar and Solar Eclipses

It is claimed by the Ahmadiyya that the Prophet (saw) prophecied that the Mahdi would appear when the moon would be eclipsed on the first day of the lunar month of Ramadan and the sun would be eclipsed in the middle of the same month.

Since this is astronomically impossible, Ahmadiyya say that first means first of 3 possible eclipse days and middle means the middle of 3 possible eclipse days. This is refuted below by David McNaughton.

Context of the Fake Ahadith

This is a summary of the Dar Qutni chapter that contains this account, not a 'hadith'

1762: Aisha(ra) said - Once Rasulullah (saw) commanded . . . . not direct quote

1763: no narration back to Rasulullah (saw)

1764: Aisha (ra) narrates how Rasulullah (saw) conducted eclipse prayer - no direct quote

1765: Ahadith comparison according to various narrators

1766: Description of Eid prayer - no direct quote

1767: ibn Abbas narrating the instruction of Rasulullah (saw) - note he is ahl-e-bait and quotes the Nabi (saw). Ahmadis say it is not necessary for ahl-e-bait to attribute to Nabi (saw)

1768: Aisha (ra) describing eclipse prayer of Rasulullah (saw) . . . not direct quote

1769: Direct quote from Prophet (saw) in inverted commas

1770: Direct quote from Prophet (saw) in inverted commas

1771: NO MENTION of Rasulullah: and 'our Mahdi' means the mahdi of Muhammad bin Ali - whose words are quoted.

1772: Direct quote from Prophet (saw).

1771 is the 'hadith' of the eclipses. Can anyone imagine Rasulullah (saw) saying 'our Mahdi'. And this chapter is about eclipses under the context of Eids. That it will be an astronomical anomaly - against the laws of astronomy - and no such event has happened in the billions of years since the creation of the heavens and the earth. This was just the opinion of Imam Baqir - Muhammad bin Ali.

Now let us look at the words:

inna - surely

li-mahdiyyina - for our Mahdi

aayatain - two signs

lam takoona - not occured

munz - since

khalaqa samawati wal ard - creation of the Heavens and the Earth

yankasiful qamaru - the moon will be eclipsed

li-awwali laila - on the first night

min ramadan - from Ramadan

wa tankasifu shams - and it will eclipse the sun

fee nisf minh - in the middle of it

wa lam takoona - and these two have not happened

munz - since

khalaqa samawati wal ard - creation of the Heavens and the Earth

Analysis by David McNaughton


by Dr David McNaughton, E-mail:

Extracted from the more detailed version at

Urdu translation by Muhammad Baig [of ICOP, UK] at Also attached.

The Holy Month of Ramadan 1311 (March/April 1894) contained both a lunar and a solar eclipse. The Ahmadiyya community attaches great importance to them, believing that they conferred unusual and special status on their leader, Mirza Ghulam Ahmed. This is based on a claim that the lunar eclipse occurred on the earliest possible date in an Islamic month - which they argue is the 13th. For them, it was also significant that (according to their reckoning) the subsequent solar eclipse took place on 28th Ramadan, supposedly occupying the 'middle' of the permissible range of dates.

Question-marks against eclipses on the 27th

The Ahmadiyyas maintain that a solar eclipse may be witnessed either during the 27th, or on the 28th, or on the 29th of a lunar month. At first glance the 27th does seem questionable, because a crescent-sighting on the following date would terminate the month incorrectly after just 28 days. Remembering that a New Moon is 'born' during a solar eclipse, its age just after sunset next day would exceed 24 hours - almost certainly making it visible somewhere in the world.

Admittedly, if people are following an Islamic calendar based on observations made just at one point, then even a 30-hour crescent will sometimes be missed, particularly if the moon lies quite a long way to the right or left of the setting sun. (Under these circumstances, incidentally, successful sightings will usually be achieved in Earth's other hemisphere).

However, by looking very carefully at the beginning of a lunar month with a solar eclipse on its 27th day, we learn about a much bigger problem. A detailed analysis of the associated time-intervals (see the "Appendix" below) - shows that this can happen only when the crescent still remains invisible 1½ days after the 'birth' of the New Moon. This is certainly possible at a time and place where the moon sets a long way to the right or left of the sun, but only if the low-level atmospheric visibility is very poor at the same time.

The delay in starting Ramadan 1311

The new crescent was not visible from Qadian (near Lahore) on 8th March 1894, so the commencement of Ramadan was postponed to the evening of March 9th. This meant that the date of the subsequent Full Moon and lunar eclipse (21st March) became 13th Ramadan rather than the 14th, thus agreeing with the Islamic date required by the Ahmadis.

At Qadian at dusk on 8th March 1894, the moon was situated 10 degrees above the sun. In hazy weather, it would admittedly have then been impossible to spot the New Moon. However, the crescent would probably have been observed from high-altitude stations to the north or east of Qadian - where the air is always thinner, less dusty, and drier.

The associated problem of 28-day months

At Qadian, it would have been sounder to commence Ramadan 1311 at dusk on 8th March. Whenever haze prevents identification of a New Moon, as a general rule it is advisable to inquire whether it was detected elsewhere, and if so, to begin the new month immediately. This is because the sky could become quite clean at the end of the same month, revealing the crescent perhaps only 16 to 20 hours after it was 'born'. And if the first day of that month had been 'lost', it might then have to finish incorrectly after just 28 days. Two (or even three) successive 29-day months can occur naturally: in those circumstances an (obviously mistaken) 28-day Islamic month is possible with the later one - if its start is postponed due to local bad weather.

Dhu al-Hijjah 1411 (13th June to 12th July 1991) at 25º South, 65º East - provides a specific example of that problem, as discussed in URL

Lunar eclipses on the 12th of a month

Rules for determining Islamic calendars vary according to community and nationality; one cannot be dogmatic as to whether any particular system is "right" or "wrong". However, it is essential to be consistent - always retaining exactly the same criteria through the entire year (otherwise 28-day and 31-day months will occasionally be experienced).

If the Ahmadiyyas wish to base their decisions on observations made just at a specified point (like Qadian, where haze may hide a young moon), then lunar eclipses will sometimes be witnessed on the 12th of an Islamic month. This is because:

  • (i) the start of a month can occasionally be delayed till 2½ days after the 'birth' of the New Moon (if crescent identification is impossible after 1½ days, as discussed earlier), and
  • (ii) because it is possible for the interval between New and Full Moon to be less than 14 days.

Here are two examples of a late start to an Islamic month resulting in an eclipse falling on its 12th day:

At 42º South, 50º East on 8th February 2008 at dusk, the crescent would not have been visible in the presence of thick low-level haze. A locally determined new month at this particular place (at sea level) could quite easily not have commenced until the evening of 9th February. In that case, the subsequent lunar eclipse on 21st February 2008 would have been observed there (before dawn) on the 12th of the Islamic month.

Poor atmospheric visibility would also have obscured the Ahmadi Ramadan crescent on 8th March 1894 at 40º South, 120º West. If based on observations made just at this point, the new month would then have had to wait till the evening of the 9th - as at Qadian. In the southeast Pacific, however, the subsequent lunar eclipse on 21st March would have been witnessed in the early hours of the morning - which in that region was still 12th Ramadan. (Less haze would be necessary for a postponement to March 9th if we moved slightly further away from the equator, say to 45º South, 120º West).

There are instances in the past, when the lunar crescent remained hidden in haze with a sun-moon separation and configuration similar to those above. (Indeed, cases like that are cited in articles written by Bradley Schaefer in the USA, and by Mohammad Ilyas in Malaysia).


Weather conditions favouring lunar eclipses on the 12th are not too different from those necessary to obtain solar eclipses on the 27th of an Islamic month, namely haze which is dense enough to disguise a 1½-day crescent.

Thus, the Ahmadiyyas must either accept that eclipses may occur on the 12th of a lunar month as well as on the 27th - or else they must regard eclipses as impossible on both those Islamic dates. Whichever choice is made, requires revision of their thesis.


To enable an eclipse to occur on the 27th of an Islamic month, that month would need to have its commencement delayed until the moon's age was 2½ days. It is easy to prove that.

Let us assume that (at the location being considered) the eclipse is observed near the end of the 27th date in the lunar month, say 26.95 days after that month commenced (which is always at dusk).

The average length of a lunar circuit is just over 29.5 days, so we will use this figure for our illustration. [In fact, it would really be better to assume a higher value. It certainly cannot be significantly less - see below**.] This 29.5-day period is then the interval between the 'birth' of the two New Moons marking - or associated with - the beginning and end of the month. Remember too that a New Moon is always 'born' during a solar eclipse.

Subtracting the first period from the second gives

29.5 - 26.95 = 2.55 days,

i.e. this particular Islamic month must have started 2.55 days after the birth of its New Moon. This means that the crescent could not have been sighted at dusk on the evening just before that month commenced; i.e. the moon still remained invisible 1.55 days after its 'birth'.

- - - -

** A study of the patterns of the moon's behaviour demonstrates that a faster than normal lunar circuit (which can be as short as 29.27 days) is not likely to be associated with this phenomenon of an eclipse on Day 27. That is because its comparatively high speed of migration makes it easier to spot the crescent after 1½ days.

An article in Urdu that explains the anomalies in the theory: