Several chemists in the 1860s noticed patterns among the known elements. What set Dmitri Mendeleev apart was the risk he took. When the pattern called for an element nobody knew about, he left a gap in his table, and then described what would eventually fill it.
Trusting the pattern
Mendeleev published his periodic table in 1869, arranging the 63 known elements mostly by atomic mass and grouping those with similar chemistry. Where no known element fit, he refused to force one in.
In 1871 he went further and predicted the properties of three missing elements, based on their neighbours in the table. He called them eka-aluminium, eka-boron and eka-silicon, using the Sanskrit word for “one” to mean one place beyond.
The predictions come true
- Gallium (1875): found by Paul-Émile Lecoq de Boisbaudran in zinc ore. It matched eka-aluminium.
- Scandium (1879): found by Lars Fredrik Nilson in Uppsala. His colleague Per Teodor Cleve recognised it as eka-boron.
- Germanium (1886): found by Clemens Winkler. It matched eka-silicon closely, as the table below shows.
| Property | Predicted for eka-silicon | Measured for germanium |
|---|---|---|
| Atomic mass | 72 | 72.6 |
| Density | 5.5 g/cm³ | 5.3 g/cm³ |
| Oxide | EsO₂ | GeO₂ |
| Chloride | EsCl₄, boils below 100 °C | GeCl₄, boils at 87 °C |
Gallium produced the best story. Lecoq de Boisbaudran measured its density at 4.7 g/cm³, and Mendeleev, who had never seen a sample, wrote to say it should be close to 6. After purifying his sample and measuring again, Lecoq de Boisbaudran got 5.9.
Where he was wrong
Not everything worked out. Mendeleev swapped tellurium and iodine, against their atomic masses, to keep their chemical families together, and could not explain why that was necessary. He also had no place for the noble gases, which were unknown at the time, and he speculated about elements lighter than hydrogen that don’t exist.
The missing explanation
The answer to the tellurium problem came in 1913, when Henry Moseley used X-rays to show that each element has a whole number that rises one step at a time: its atomic number, the count of protons in its nucleus. Ordered by that number, tellurium and iodine fall into place with no need for a swap.
Moseley’s work also showed exactly which gaps remained. One of them, the eka-manganese Mendeleev had also anticipated, was filled in 1937 by technetium, the first element made artificially. His table outlived every assumption he built it on, because the patterns in it were real.