The atom was proposed twenty-four centuries before anyone saw evidence for it. What happened in between is one of the best worked examples in science of a model being repeatedly demolished and rebuilt — not because the earlier chemists were careless, but because each new instrument revealed something the previous model could not accommodate.

Here is that sequence, and what each step actually contributed.

Democritus, c. 400 BC: the atom as an argument

Leucippus and his student Democritus asked what happens if you keep halving a piece of matter. Either you divide forever, or you reach something indivisible. They chose the second answer and called it atomos — uncuttable.

This was philosophy, not chemistry. There was no experiment behind it and no way to test it. Aristotle rejected it in favour of the four elements, and because Aristotle's authority held for the better part of two thousand years, so did his rejection. The atom stayed a minority opinion until quantitative chemistry arrived to give it a job.

Dalton, 1803: the atom becomes measurable

John Dalton's contribution was to make atoms do explanatory work. Studying how elements combine, he noticed that they do so in fixed, small, whole-number ratios by mass — the law of multiple proportions. Carbon and oxygen form carbon monoxide and carbon dioxide, and the oxygen masses are in an exact 1:2 ratio. Never 1:1.7.

That is very hard to explain if matter is continuous, and trivial if matter comes in discrete units. Dalton's model:

• Elements consist of indivisible atoms.
• All atoms of an element are identical in mass and properties.
• Compounds form when atoms combine in fixed whole-number ratios.
• Chemical reactions rearrange atoms but never create or destroy them.

Two of those four are now known to be wrong — atoms are divisible, and isotopes mean atoms of an element are not identical in mass. It didn't matter. Dalton turned the atom from a philosophical position into a quantitative tool, and chemistry as a discipline followed.

Thomson, 1897: the atom acquires parts

J. J. Thomson was studying cathode rays — beams in evacuated tubes. He showed the beam bent under electric and magnetic fields, always in the direction indicating negative charge, and always with the same charge-to-mass ratio no matter which metal produced it or which gas filled the tube.

The implication was unavoidable: there is a negatively charged particle far lighter than any atom, and it is present in all of them. The electron.

Which broke Dalton's first postulate. If atoms contain smaller things, they are not indivisible. Thomson proposed the plum pudding model — electrons embedded in a diffuse sphere of positive charge, like fruit in a cake. It accounted for neutral atoms containing negative parts, and it was the best available answer for fourteen years.

Rutherford, 1911: the atom turns out to be mostly nothing

The experiment that ended the plum pudding is the most elegant in this history. Rutherford, with Geiger and Marsden, fired alpha particles at gold foil only a few hundred atoms thick, and recorded where they landed.

The plum pudding predicted every particle would pass through with a slight deflection, since diffuse positive charge cannot exert much force. Almost all did. But roughly one in eight thousand bounced back toward the source.

Rutherford's remark is the best description of a falsified model anyone has written: it was as if you had fired a fifteen-inch shell at tissue paper and it had come back and hit you.

The only structure that produces that result is one where almost all the mass and all the positive charge sit in a minute central volume. The nuclear model: a dense nucleus, electrons somewhere outside it, and overwhelmingly empty space between. The nucleus occupies roughly one part in a trillion of the atom's volume.

It had a fatal problem. Classical physics says an orbiting charge radiates energy continuously. A Rutherford electron should spiral into the nucleus in about ten picoseconds. Matter should not exist.

Bohr, 1913: energy comes in steps

Niels Bohr resolved it by refusing the classical requirement. Electrons, he proposed, occupy fixed energy levels and cannot radiate while in one. They move between levels only by absorbing or emitting a photon of exactly the energy difference.

The evidence was already sitting in the laboratory. Heat hydrogen and it emits light at sharp, specific wavelengths — not a continuous spectrum. Bohr's model predicted hydrogen's emission lines with precision.

It worked beautifully for hydrogen and failed for everything with more than one electron. But the central idea — quantised energy levels — survived every revision that followed and is why the periodic table has the shape it does.

Schrödinger and Heisenberg, 1926: the orbit becomes a probability

The quantum mechanical model replaced the electron's path with a probability distribution. Heisenberg's uncertainty principle established that position and momentum cannot both be known precisely, so a defined orbit is not merely unknown but meaningless. Schrödinger's wave equation describes instead where an electron is likely to be found.

Those probability regions are orbitals — the s, p, d and f shapes that determine bonding geometry, and therefore molecular shape, and therefore essentially all of chemistry. This is the model still in use.

Chadwick completed the nucleus in 1932 with the neutron, explaining isotopes: same element, same proton count, different mass. Dalton's second postulate finally fell.

What the sequence is actually about

Each model was overturned by an instrument, not an argument. Cathode ray tubes found the electron. Alpha scattering found the nucleus. Spectroscopy forced quantisation. The philosophy of the atom went nowhere for two thousand years; the measurement of it took a century and a quarter.

Which is the through-line worth keeping. Chemistry advanced when it could measure — and the objects that did the measuring, the tubes and flasks and spectrometers, are as much the story as the people. That is a large part of why laboratory glassware has the forms it does: each vessel is a solved problem in containing, heating, or measuring something precisely.

It is also why we make what we make. The Flask Vessel takes the Erlenmeyer geometry — a form designed in 1860 to swirl a solution without spilling it — and rebuilds it in stainless steel for daily use. If the instruments of measurement appeal to you as objects, our chemistry gifts and science drinkware are built on that idea, and the Erlenmeyer flask bottle guide explains the form in more detail.

The models at a glance

• c. 400 BC — Democritus: matter is made of indivisible atomos. Philosophical, untested.
• 1803 — Dalton: atoms are indivisible, identical within an element, and combine in fixed ratios. Quantitative at last.
• 1897 — Thomson: electrons discovered; plum pudding model. Atoms have parts.
• 1911 — Rutherford: gold foil experiment; dense nucleus, mostly empty space.
• 1913 — Bohr: quantised energy levels explain emission spectra.
• 1926 — Schrödinger & Heisenberg: orbitals replace orbits; probability replaces position.
• 1932 — Chadwick: the neutron explains isotopes.