The four minutes I chose not to drive to

3 September 2026astronomytravelnotesvideo

When I wrote up the eclipse trip I said that 93% is a strange and beautiful lighting effect and that 100% is apparently something else entirely. I was being slightly defensive. I have now had three weeks to think about it, and I want to be honest about how we ended up on that ridge rather than inside the path.

The decision

We did not book early enough. By March, when I wrote the plan down, everything inside the path of totality within reach was either gone or priced for people with very different budgets. We found a room in a village a little over an hour's drive south of the edge of the path, and the plan was to drive north on the afternoon itself.

Then, a few days before, three things happened at once. The forecasts started disagreeing about cloud along the stretch of path we had picked. Friends who had driven to eclipses before were predicting heavy traffic on the roads into the path. And the eclipse itself was low in the western sky in the evening, which meant the place you chose had to have a clear horizon in exactly the right direction, not just a clear sky.

I did the sums. Totality where we would have gone was a bit under a minute and three quarters. The drive was an hour each way in good conditions and possibly three or four hours in bad ones, with the risk of being stuck on a road with the Sun behind a hill at the critical moment. Against that: a ridge near the village with a perfect western view, a short walk from the room, and 93%.

We stayed. In March I had written that a deep partial in sunshine beats a total behind cloud, and I held on to that sentence a little too gratefully. I told myself it was the sensible choice, and in terms of logistics it probably was. The cloud bank that sat over the next valley all afternoon did not, in the end, drift anywhere useful, so I will never know whether the drive would have paid off.

What I missed

Since getting home I have read accounts from people who were inside the path that evening, a few of them from places less than a hundred kilometres from our ridge. They all say the same things, which is what convinced me they are not exaggerating. The light does not dim gradually and then a bit more; in the last few seconds it collapses. The temperature drops sharply. And then the corona appears, and nobody who has seen it seems to be able to describe it without sounding faintly foolish.

So last weekend I sat down with a cup of tea and watched twenty minutes of NASA's telescope feed from the total eclipse over Mexico in April 2024. This is not my footage, obviously; it is NASA's, filmed from inside the path in Mexico, where totality lasted over four minutes, more than twice as long as anything Spain got this year.

If you are short of time, skip to roughly the middle, but I would recommend watching it from the start, because the slow approach is the point. For several minutes you see a thinner and thinner crescent, which is what I saw through my filter on the ridge. Then the last bead of sunlight breaks up along the Moon's edge into separate points, as light slips through valleys on the lunar limb, and there is one bright point left on one side of a ring of light. That is the diamond ring, and the moment it goes, the filter comes off and the corona is simply there.

What surprised me most was the pink. Around the black disc there are small bright pink-red flames: prominences, loops of hot gas held up by the Sun's magnetic field, which you only see during totality or through a specialised solar telescope. In the 2024 feed some are large enough to pick out at a glance. I had seen photographs of them, of course, but I had not appreciated how vivid they are against the white of the corona, or how quickly they vanish when the first sliver of the Sun's surface returns at the end.

Then the second diamond ring, on the opposite side, and it is over. Four minutes looks short on a video, and everyone who has been there says it feels much shorter in person.

What I think now

I do not regret the ridge as such. It was a good evening. The light really was extraordinary, the projected crescents on the card really were sharper than anything on a screen, and the baguette from the village bakery was the best bread I have eaten this year. If the choice was 93% or nothing, I would take 93% again.

But it was not a choice between 93% and nothing. It was a choice between 93% and a reasonable chance of totality, and I picked the option with less risk and much less reward. Watching the feed made that clear in a way that reading about it never did. A partial eclipse, however deep, is a change in the daylight. Totality is a different event, and 93% is not 93% of it.

The lessons, written down so I cannot pretend later that I did not know them:

1. Book early, and book inside the path, even if it means a worse room.
2. Plan to be mobile on the day, but start from a place where staying put is still totality.
3. Check the horizon in the direction of the Sun at the right time, not just the cloud forecast.
4. Bring the filter and the card, because the partial phases are worth having anyway.
5. Buy the bread at six in the morning, because it really will be gone by eleven.

The next total eclipse within reasonable travelling distance is on 2 August 2027, crossing the far south of Spain and North Africa, with totality of several minutes in places. I have opened a document called "2027" and written the first item on the list at the top of it. It is underlined twice.