The Pantheon Oculus Light in Autumn: Where Rome's Sunbeam Goes When It Leaves the Floor

The rotunda is as wide as it is tall, about 43.3 metres either way. The drum, which is the cylindrical wall you stand inside, stops at 21.65 m, exactly half the height, and the dome curves up from there. That single proportion decides everything about the light. The measurements themselves, and the places where surveys disagree about them, are set out on our page about the dome and the oculus.

A beam coming through the opening falls at whatever angle the sun happens to be at. To clear the drum and reach the pavement, the sun has to stand higher than 63.4 degrees. Rome is at 41.9 degrees north, so that happens for about a hundred days a year. Our figures for this year give 2 May to 10 August, which is 101 of them, give or take a day at either end depending on the year.

Below 63.4 degrees the disc lands on the drum wall instead. Below 45 degrees it does not reach the wall at all and stops on the dome, among the coffers. Those two thresholds are the whole of autumn. In the middle of today the sun over Rome gets to 47.8 degrees, so the light is on the wall, well above the door. Next Wednesday it manages 45.1, about as close to the join between wall and dome as it ever comes. After that it is on the curve until March.

The times below are the moment the beam crosses the axis of the building, which is neither noon nor solar noon. Two separate things push it about. Italian clocks are set to a meridian well east of Rome, which already puts solar noon about ten minutes late, and summer time adds an hour on top until 25 October. Then there is the building, turned roughly four and a half degrees west of north, so the beam sweeps over its axis about ten minutes before the sun reaches due south.

Read the two rows either side of the clock change. The sun goes on doing what it has done all autumn, shedding a couple of degrees a week. The clock drops a whole hour in one night, and for most visitors that is the more disorienting of the two.

On 30 September the centre of the disc sits at 21.6 m, a few centimetres under the springing line where the drum ends and the curve begins. On 1 October it is on the dome, and there it stays for five months. If you want to see the beam strike something flat and vertical rather than a curve, this week and next are what remains of it.

After that it climbs. By the solstice the centre of the lit patch is 35.8 m up and only 16.4 m out from the middle of the room, which is the highest and the tightest it gets all year. That is the winter version, and it is not a consolation prize. The coffers throw shadows that shift while you stand there, and the patch slides along the curve instead of sitting still on the floor.

For the light, the useful window in autumn is the forty minutes or so around the axis crossing. Until 25 October that means half past twelve to one o'clock. After it, a quarter past eleven to noon. Outside that window the disc rides higher and stretches out of shape: at nine this morning it was 37.6 m up the dome, and at six this evening it will be at 41.6 m, faint and spread across several coffers just under the opening.

Here is the part no seller of tickets wants to write down. The hour with the best light is the hour with the worst room. Entry runs in timed bands with a ceiling of 1,200 people in each, and it is the middle of the day that fills first. Nine in the morning gives you an almost empty floor and a pale smear near the top of the dome. One o'clock gives you a bright, compact disc and several hundred people between you and the wall it has landed on. Nobody can sell you both, and our page on the best time to visit has the rest of that trade-off, including the days when the room is unavoidably full.

On 24 October the sun sets over Rome at 18:15. On the 25th it sets at 17:13. Nothing has happened in the sky. The clocks went back at three in the morning and took the evening with them.

The monument keeps the same hours all year, 9:00 to 19:00, with the last admission at 18:30, as the Ministry of Culture publishes on its own page for the building. In September that still means closing in daylight, barely: the sun sets at 19:05 today and at 18:58 on 27 September, and from that Sunday until the spring the doors shut after dark. Once the clocks change, anything from about a quarter past five onwards is a night visit, which alters what you are paying for. No beam, no shadows travelling across the coffers, just the shape of the place and the hole in the roof. The calendar version of all this is on opening hours and closures, and if you want daylight outdoors afterwards, the sunset walks around the quarter begin about when the monument stops being interesting.

The September equinox falls at 02:05 Rome time today, 23 September. Inside the building nothing marks it. The disc is 19.4 m up the entrance wall, it will be slightly higher tomorrow, and it was slightly lower yesterday. There is no alignment and no moment to turn up for. An equinox is a fact about the Earth's axis, not about this room.

What the arithmetic does produce is a mirror. Today's position on the wall is within a metre of where the disc sits in the third week of March, and the same symmetry holds either side of the solstice: on 21 April the centre of the disc crosses the great doorway at 5.9 m above the floor, and on 21 August at 5.6 m. That symmetry is the strongest objection to the best story told about this building.

In 2011 Robert Hannah and Giulio Magli argued in Numen that the April alignment was designed, with an emperor walking in through the sunlight on Rome's traditional birthday. The beam itself is real and anyone can reproduce it; we just have. Whether it was intended is a hypothesis, and the authors are careful about that: they state plainly that they are not claiming the building was made to measure the sun. Eugenio La Rocca reached a similar reading of April independently, then pulled back from it, calling sundial interpretations of the rotunda more than the evidence will carry. The standard Cambridge volume on the monument leaves the theory alone. And the sky cannot break the tie, because the sun stands at the same height over Rome in the third week of August as it does on 21 April. The full month by month calculation is on where the light falls.

This is the season the question starts to matter, and the answer is not the one you will hear in the square. Warm air rising through the opening does sometimes break up a light drizzle. It does not keep the weather out. The most direct source on that is the basilica that holds services in the building.

Water lands on the marble and drains away underneath it. The historical argument is blunter than the physics. Giacomo della Porta in 1591, Alexander VII in 1660 and James Gibbs in the eighteenth century all worked on schemes to close the oculus, and in 1880 a proposal for a tomb in the middle of the floor was refused partly because the ancient drains below would not have coped. Three centuries of expensive worry about plumbing is not the signature of a building that stays dry.