Dmitri Mendeleev (1834–1907) was professor of chemistry at St Petersburg University in early 1869, working on a textbook (Osnovy khimii, “Principles of Chemistry”) and substantially substantively trying to organise the known chemical elements into a coherent presentational order. He had cards listing each element’s name, atomic weight, and chemical properties laid out on his desk through January and February 1869.
The story he told his biographer Aleksandr Inostrantsev decades later: on the morning of 17 February 1869 (5 March in the Gregorian calendar), after three days of unsuccessful waking work, he fell asleep at his desk. He dreamed of a table in which all the elements fell into their proper places. On waking he wrote out the arrangement immediately; the dream version required only one small correction during subsequent revision.
The story is substantially probably embellished. Mendeleev’s working papers from the substantial relevant week survive; they show a continuous evolution of the table through January and February rather than a sudden overnight insight. Modern historians of chemistry treat the dream narrative as substantively a substantively retroactive simplification.
What the table did
Mendeleev arranged the 63 then-known elements by ascending atomic weight, broken into rows that brought elements with similar chemical properties (alkali metals, halogens, etc.) into vertical columns. The arrangement worked. It also had gaps — places where the column-pattern required an element that had not yet been discovered.
Mendeleev predicted the properties of three then-unknown elements that would have to fill specific gaps: eka-aluminium (predicted atomic weight ~68, similar to aluminium), eka-boron (~44, similar to boron), and eka-silicon (~72, similar to silicon).
All three were substantively discovered within seventeen years. Gallium (1875, Paul Émile Lecoq de Boisbaudran) matched Mendeleev’s eka-aluminium predictions with substantively striking accuracy — when Boisbaudran’s first reported atomic weight differed from Mendeleev’s prediction, Mendeleev wrote to Paris substantively suggesting Boisbaudran had got his measurement wrong, and Boisbaudran’s revised measurement agreed with Mendeleev. Scandium (1879, Lars Fredrik Nilson at Uppsala) matched eka-boron. Germanium (1886, Clemens Winkler at Freiberg) matched eka-silicon.
The predictions were what substantively converted the broader European chemical community to the periodic-table framework. Without them the table would have been substantively only a substantively useful pedagogical mnemonic; with them it became substantively the organising principle of modern chemistry.
Mendeleev died of influenza at St Petersburg in February 1907, aged 72. He had been nominated for the Nobel Prize in chemistry three times. He never won it; the 1906 committee that should have given it to him substantively voted instead for the French chemist Henri Moissan, in what is widely substantively considered the Nobel committee’s most substantively significant 20th-century scientific oversight.