The Short Answer
Every month the Moon disappears twice over: as an old crescent it rises a little before the Sun and thins out morning by morning, then it passes the Sun at the new Moon, then it comes back as the young crescent in the evening. We worked out every new Moon of 2027 at four cities, with the same rule for mornings and evenings (Yallop's criterion, and a dark enough sky with the Moon high enough to search in). At Jakarta, Mecca and Cape Town the old crescent could be found two and three mornings before the new Moon in every month of the year, but on the very last morning only 4 or 5 times in 13. The gap from the last morning crescent to the first evening one was usually two and a half days: see it at dawn, miss it the whole of the next day, and find the new crescent on the evening after that. The shortest gap was a day and a half, the longest three and a half. London has longer gaps, up to four and a half days, because its winter mornings and its autumn evenings hold the Moon low. All of this assumes a clear sky and a flat horizon.
The old crescent, morning by morning
After the full Moon, the Moon rises about 50 minutes later each day (we explain why in the day the Moon does not rise). By the last quarter it rises around midnight; a week later it rises only an hour or two before the Sun, as a thin crescent low in the east, with its lit side facing down towards the Sun that is about to rise. Each morning it rises closer to sunrise and is thinner, until it is lost in the dawn.
At Mecca, around the new Moon of 6 February 2027, it goes like this (local time, clear sky, flat horizon):
| Morning | Moonrise | Sunrise | Moon at sunrise | Lit | Hours to the new Moon | Verdict |
|---|---|---|---|---|---|---|
| 3 February | 04:38 | 06:58 | 24.4° up | 10.7% | 84 | Zone A, easy |
| 4 February | 05:24 | 06:57 | 17.2° up | 5.6% | 60 | Zone A, easy |
| 5 February | 06:05 | 06:57 | 9.7° up | 2.1% | 36 | Zone A, search 06:23 to 06:48 |
| 6 February | 13 minutes before the Sun | 06:56 | 2.2° up | 0.2% | 12 | Zone E, no time to search |
So the last chance that month is the morning of 5 February, 36 hours before the new Moon. On 6 February the Moon is still there, but it rises only 13 minutes before the Sun, a sliver 0.2 per cent lit, and it is never high enough while the sky is dark enough. The new Moon comes at 18:56 Mecca time on 6 February, and the young crescent can first be found on the evening of 7 February, 23 hours after it (zone B, search 18:23 to 18:42). The sky has no crescent anyone could find for two and a half days in between.
How often the last mornings can be found
We did the same for every one of the 13 new Moons of 2027 at four cities: Jakarta (6° S), Mecca (21° N), Cape Town (34° S) and London (52° N). A morning counts as findable when two things hold, the same two that the evening plan uses:
- Yallop's zone is A to D at the morning's best time. Zones A and B mean it can be seen by eye (B in perfect conditions); in C binoculars may be needed to find it first, and in D they will be. Zone E means it is very unlikely to be seen at all.
- There is a window to search in: a time when the Moon is more than 3 degrees up while the Sun is still more than 2 degrees below the horizon. A crescent with a good zone but no window, because it rises too close to the Sun, is what the app calls "very hard": low in bright twilight.
Number of the 13 months in 2027 in which each morning and evening could be found:
| Jakarta | Mecca | Cape Town | London | |
|---|---|---|---|---|
| Third to last sunrise before the new Moon | 13 | 13 | 13 | 12 |
| Second to last sunrise | 13 | 13 | 13 | 9 |
| Last sunrise before the new Moon | 5 | 5 | 4 | 3 |
| New Moon | ||||
| First sunset after the new Moon | 3 | 3 | 4 | 1 |
| Second sunset | 13 | 13 | 10 | 8 |
| Third sunset | 13 | 13 | 13 | 10 |
The pattern is the same on both sides of the new Moon: one to three days away from it the crescent is a fair chance almost everywhere, and on the morning or evening nearest the new Moon it is a long shot. On the last morning, a crescent that could be seen by eye (zone A or B with a window) came only 3 times in Jakarta, 4 in Mecca, 2 in Cape Town and once in London.
How close to the new Moon the last findable morning came, and how soon after it the first findable evening:
| City | Last findable morning: hours before the new Moon | First findable evening: hours after the new Moon |
|---|---|---|
| Jakarta | 15.2 to 41.0 | 17.2 to 43.3 |
| Mecca | 14.6 to 40.4 | 15.8 to 45.9 |
| Cape Town | 18.8 to 40.7 | 14.2 to 55.7 |
| London | 15.9 to 80.3 | 18.9 to 87.0 |
The shortest times are the rare, hard cases: Mecca's 14.6 hours is the morning of 31 August 2027, zone D, which means binoculars or a telescope, and is a prediction of the criterion, not a sighting. As with the evening crescent, the age of the Moon is a poor guide on its own: what counts is how far the Moon is from the Sun and how high it stands when the sky is dark enough.
The moonless gap
From the sunrise of the last findable morning to the sunset of the first findable evening there is no crescent anyone could find. Because the gap runs from a sunrise to a sunset, it always comes out close to a whole number of days and a half.
| City | Shortest | Longest | Average | 1.5 days | 2.5 days | 3.5 days | 4.5 days |
|---|---|---|---|---|---|---|---|
| Jakarta | 1.5 | 3.5 | 2.4 | 2 | 10 | 1 | 0 |
| Mecca | 1.5 | 3.5 | 2.5 | 2 | 9 | 2 | 0 |
| Cape Town | 1.5 | 3.6 | 2.7 | 1 | 8 | 4 | 0 |
| London | 1.7 | 4.5 | 3.6 | 1 | 3 | 2 | 6 |
(London's row has 12 months, not 13: after the new Moon of 31 August no evening of the first four had a window, as the next section explains.)
The usual gap at the three lower-latitude cities is two and a half days. In practice that means one morning with the last crescent, one whole day with none, and the new crescent on the evening after. The best months shrink it to a day and a half: the old crescent at dawn on one day, the new one at dusk the next. In 2027 that happened twice each at Jakarta (31 August to 1 September, 29 to 30 October) and Mecca (6 to 7 April, 4 to 5 June), once at Cape Town (29 to 30 September) and once at London (4 to 5 June). No morning and evening of the same day ever both had a crescent: the hours before and after the new Moon would have to add up to one day's daylight, about 12, while the closest morning and the closest evening of the whole year at these cities add up to more than 28.
Why London's gaps are longer
At London the gap was about four and a half days in six of the 13 months of 2027, and the reasons are different in winter and in autumn:
| New Moon (London time) | Last findable morning | First findable evening | Gap (days) |
|---|---|---|---|
| 7 January, 20:24 | 5 January (60 h before) | 9 January (44 h after) | 4.3 |
| 6 February, 15:56 | 3 February (80 h before) | 7 February (25 h after) | 4.4 |
| 8 March, 09:29 | 6 March (51 h before) | 9 March (32 h after) | 3.5 |
| 7 April, 00:51 | 5 April (42 h before) | 7 April (19 h after) | 2.6 |
| 6 May, 11:59 | 5 May (31 h before) | 7 May (33 h after) | 2.6 |
| 4 June, 20:40 | 4 June (16 h before, zone D) | 5 June (25 h after) | 1.7 |
| 4 July, 04:02 | 3 July (23 h before) | 5 July (41 h after) | 2.7 |
| 2 August, 11:05 | 1 August (30 h before) | 4 August (58 h after) | 3.6 |
| 31 August, 18:41 | 30 August (37 h before) | none in the first four evenings | |
| 30 September, 03:36 | 29 September (21 h before, zone C) | 3 October (87 h after) | 4.5 |
| 29 October, 14:36 | 28 October (31 h before) | 1 November (75 h after) | 4.4 |
| 28 November, 03:24 | 26 November (44 h before) | 30 November (61 h after) | 4.3 |
| 27 December, 20:12 | 25 December (60 h before) | 29 December (44 h after) | 4.3 |
From December to March it is the mornings that fail: the last findable morning comes 51 to 80 hours before the new Moon. From August to November it is the evenings: the first findable evening comes 58 to 87 hours after it, and in September none of the first four evenings has one. In between, from April to July, both sides are good and the gap is as short as anywhere.
The reason is the angle at which the Moon's path meets the horizon. The Moon travels close to the ecliptic, the Sun's yearly path, and at a northern mid latitude the ecliptic stands steeply over the western horizon on spring evenings and over the eastern horizon on autumn mornings, but lies low along the horizon on autumn evenings and on late winter and spring mornings. When the path is steep, a crescent a day from the Sun is already well above the horizon at sunrise or sunset; when it is shallow, the same crescent hugs the horizon and is lost in the twilight. The Moon's own path tilts about 5 degrees either side of the ecliptic, which can help or hurt in a given month. Why the Moon's path across the sky changes and the crescent's tilt by latitude and season show this on the evening side.
The most striking case is September 2027. After the new Moon of 31 August, Yallop's zone for London is B on 2 September and A on 3 September, but the Moon is only 1.6 and 3.9 degrees up at sunset and sinks below 3 degrees before the sky is dark enough, so neither evening has a window. The first evening with one is 4 September, 97 hours after the new Moon, by which time the crescent is 21 per cent lit and past the app's four-day crescent search. From the last morning crescent on 30 August that is five and a half days without a crescent to find.
Cape Town, at 34° S, shows the mirror of London's autumn evenings: its slowest first evenings came from February to June, its own autumn, at 41 to 56 hours after the new Moon; in the other months they came within 39 hours. Its mornings were much like Mecca's all year.
Planning a morning
The Observe page and the Android app (in testing) have a Morning: old crescent view beside the evening one. It lists the mornings before the next new Moon at your place, marks the last chance, and for each one gives the zone, the window to search in, the best moment (sunrise minus four ninths of the time from moonrise to sunrise, the mirror of Yallop's evening rule) and a sketch of where to look. A few things are different from an evening:
- Arrive in the dark. The Moon rises before you can see the horizon well, so be in place early and find the eastern horizon first. The window is short: on 5 February 2027 at Mecca it is 25 minutes, ending 9 minutes before sunrise.
- Look low in the east, near where the Sun will rise. The crescent stands above and to one side of that point, with its lit edge facing down towards the Sun; the sketch in Observe gives its direction for your place.
- Put the binoculars away before sunrise. In the morning the Sun comes up right next to the crescent. Never point binoculars, a telescope or a camera at or near the Sun: it can blind you in an instant.
- Use the two mornings before the last. In 2027 they were findable in every month at Jakarta, Mecca and Cape Town. The last morning before the new Moon is a challenge, not a plan.
- Hills matter more at dawn than you expect. A low eastern horizon is as important as a low western one in the evening; Observe puts your own skyline into the times. See why altitude and terrain matter.
The old crescent is also when earthshine is at its best: the dark part of the disc glows faintly, lit by the nearly full Earth.
FAQ
How many days before the new Moon can you see the old crescent? At Jakarta, Mecca and Cape Town in 2027 it could be found on the second and third mornings before the new Moon in every month, and on the last morning in 4 or 5 months out of 13. At London the second to last morning worked in 9 months out of 13. On the last findable morning the new Moon was between about 15 and 41 hours away at the three lower-latitude cities.
How many days is the Moon invisible each month? From the last morning crescent to the first evening crescent was usually two and a half days at Jakarta, Mecca and Cape Town in 2027, at shortest a day and a half and at longest three and a half. At London it was up to four and a half days, and once five and a half. These are for a clear sky and a flat horizon; cloud and hills make it longer.
Can you see the old crescent and the new crescent on the same day? Not in any month of 2027 at these four cities. The best was a day and a half: the old crescent at dawn one day and the new one at dusk the next.
Is the old crescent easier to see than the new one? Not as a rule. The tables show both sides behave alike at the same distance from the new Moon. Which one is better in a given month depends on the season and your latitude: at London the morning side was worse from December to March and the evening side from August to November.
Do Yallop's zones work for the morning? Yallop's criterion was built from evening sightings. Using it for the morning, with sunrise in place of sunset and the time from moonrise to sunrise in place of the lag, is an adaptation, as the app's method note says. Odeh's criterion, applied the same way to the same mornings, disagreed with Yallop's on at most 3 of about 48 mornings at each city.
How this was worked out
For each new Moon of 2027 we took the mornings from three days before the new Moon's local date until the last sunrise before it, and the first evenings after it, at Jakarta, Mecca, Cape Town and London at sea level with a flat horizon. Each morning was run through the same morning plan the site and the app use (Yallop's q at the morning best time, the Moon up at sunrise, and the window), and each evening through the evening plan, with the same gate for both: a morning counts only before the new Moon and an evening only after it, within four days. The new Moon times are the moment the Moon and Sun share the same apparent longitude, rounded to the nearest minute. A script in the project's repository works out every number quoted here and its tests check them. Weather is not included: every figure is for a clear sky.
See Also
- A Beginner's Guide to Spotting the Crescent Moon
- The Youngest Crescent Moon Ever Seen
- The Science Behind the Crescent: Yallop and Odeh
- Why the Moon's Path Across the Sky Changes
- Earthshine: Why the Dark Part of the Crescent Glows
- Naked Eye, Binoculars or Telescope?
- Why Can't I See the Moon Tonight?
References
- Yallop, B.D. (1997). A Method for Predicting the First Sighting of the New Crescent Moon. NAO Technical Note No. 69, HM Nautical Almanac Office, Royal Greenwich Observatory.
- Odeh, M.Sh. (2004). "New Criterion for Lunar Crescent Visibility." Experimental Astronomy 18, 39 to 64.
- Cross, D. astronomy-engine, version 2.1.19, the open source astronomy library behind the positions, rising and setting times used here. https://github.com/cosinekitty/astronomy