Lunar eclipses could date the 1808 eruption, if someone in the Americas described the November 1808 one

During a total lunar eclipse the Moon is lit only by sunlight bent through Earth's atmosphere, so after a big eruption eclipses go very dark.

Correction, 24 October 2026: we first thought nobody had tried this for 1808. Stothers did. He searched more than 300 journals for eclipse colours from 1801 to 1881 and found none usable between 1805 and 1812, writing that reports are 'too sparse in this period to detect this eruption'. The 1812 eclipse he did find was normal, so the veil was gone by then.

What we add: Caldas's description of the 9 May 1808 eclipse in Bogota, orange and reddish, a normal Moon seven months before the eruption. Stothers' table has nothing between 1805 and 1812, so this fills a hole in his series. And the decisive eclipse, the total one of 3 November 1808, could be seen only from the Americas and the Pacific, which a mostly European search would miss.

Eclipses to use: 10 May 1808 total (Americas, normal per Caldas); 3 November 1808 total (Americas, Pacific); 30 April 1809 partial 87 percent (Europe, Africa, eastern Americas); 23 October 1809 partial 83 percent (Americas, Pacific, East Asia).

Searched with no result: the Bogota weekly 1808 to 1809, American Philosophical Society Transactions, American Academy Memoirs, the German astronomy journal 1809 to 1813, and by hand (Ken) Chronicling America, HathiTrust and the Gazeta de Mexico. An expert's search of 300 journals also came up empty, so this needs Spanish-American, Japanese or ship sources, not more European ones.