On March 1, 1869, Dmitri Ivanovich Mendeleev, professor of chemistry at Saint Petersburg University, presented to the Russian Chemical Society a classification table of known elements. Sixty-three elements were arranged by increasing atomic mass. Above all, three cells were left empty. Mendeleev announced that these cells corresponded to elements still unknown, whose properties he predicted. This prediction, audacious for the time, would become the decisive test of chemical periodicity.
Method: interpolation between neighbors
Mendeleev started from a strong hypothesis: an element's properties are a periodic function of its atomic mass. If a cell is empty between two known elements, the missing element should have intermediate properties. But not only: it must also be consistent with its vertical neighbors (same group).
Mendeleev applied this double interpolation — horizontal (period) and vertical (group) — to numerically estimate atomic mass, density, melting point, metallic character, and typical compounds for the three missing elements. He provisionally named them eka-boron (to be discovered below boron), eka-aluminium (below aluminium), and eka-silicon (below silicon), where "eka" means "one" in Sanskrit.
Prediction 1: eka-aluminium (1871)
Mendeleev predicted for eka-aluminium:
| Property | Predicted (1871) | Measured (1875, gallium) |
|---|---|---|
| Atomic mass | ~68 | 69.7 |
| Density | 5.9 g/cm³ | 5.91 g/cm³ |
| Melting point | low (Tm < 100 °C) | 29.8 °C |
| Chloride | Ea₂Cl₆ volatile | Ga₂Cl₆ volatile |
| Method of discovery | spectroscopy | spectroscopy |
In 1875, the Frenchman Paul-Émile Lecoq de Boisbaudran, analyzing a sphalerite sample, identified a new element by its characteristic spectral lines. He named it gallium (after Gallia, Latin for France) and published its properties. Mendeleev read the paper, noted the remarkable agreement with his predictions, and wrote to Lecoq de Boisbaudran suggesting he re-measure the density — which he had reported slightly low (4.7 in the first report). Re-measured, the correct value was 5.91 g/cm³ — exactly Mendeleev's prediction. The chemical community was stunned.
Prediction 2: eka-silicon (1871)
For eka-silicon, below silicon:
| Property | Predicted (1871) | Measured (1886, germanium) |
|---|---|---|
| Atomic mass | ~72 | 72.6 |
| Density | 5.5 g/cm³ | 5.32 g/cm³ |
| Color | dark grey | grey-white |
| Melting point | high (~800 °C) | 938 °C |
| Oxide | EsO₂, white, high Tm | GeO₂, white, Tm 1115 °C |
| Chloride | EsCl₄ liquid volatile, d ~1.9 | GeCl₄ liquid, d = 1.87 g/cm³, Tb 86 °C |
The German Clemens Winkler, in 1886, isolated a new element from argyrodite (silver-germanium sulfide). He published its properties and named the element germanium in honor of his country. The match with Mendeleev's predictions was even more striking than for gallium. Winkler publicly acknowledged the debt: "This correspondence proves how exact the periodic law is."
Prediction 3: eka-boron (1871)
For eka-boron:
| Property | Predicted (1871) | Measured (1879, scandium) |
|---|---|---|
| Atomic mass | ~44 | 44.96 |
| Density (oxide) | ~3.5 g/cm³ | 3.86 g/cm³ |
| Acid character | strongly basic | strongly basic |
The Swede Lars Fredrik Nilson, in 1879, isolated a new element from the Scandinavian minerals euxenite and gadolinite. He named it scandium after Scandinavia. Cleve, his colleague, made the connection with the eka-boron predicted eight years earlier and published an acknowledgment.
Score: three for three
In 15 years, the three empty cells of the 1869 table were filled, exactly as predicted. More importantly: numerical values for atomic mass, density, melting point matched within a few percent. For the 19th-century scientific community, this was proof that an underlying law governed the order of elements — even if the physical cause of that law (quantum mechanics) wouldn't be elucidated for another 60 years.
What Mendeleev didn't predict
Mendeleev was also wrong, and it's worth noting:
- He did not predict the existence of noble gases. When William Ramsay discovered them (1894-1898), Mendeleev initially refused to believe — his table had no slot for them. He eventually accepted them and added a new "group 0."
- He misplaced several lanthanides, which he couldn't tell apart from the d-subshell at the time.
- He argued for years against the atomic mass of tellurium (127.6 > iodine 126.9), which seemed to contradict the by-mass ordering — he was right on the periodicity (Te before I), but the correct criterion (atomic number, not atomic mass) wouldn't be proposed until 1913 by Henry Moseley.
The legacy
The success of Mendeleev's predictions made the periodic table the archetype of a predictive law in science — a graphic representation that holds enough information to anticipate the unknown. This property makes it a unique pedagogical object, and it's no accident that almost every chemistry textbook in 2025 opens with it.
Mendeleev died in 1907, nine years before Moseley definitively reordered elements by atomic number, providing the modern table's theoretical basis. At his death, the table had 86 elements. Today it has 118. But its 1869 structure — columns by periodicity, prediction by interpolation — has remained intact. It's one of the most enduring constructions in the history of science.