The most successful piece of predictive chemistry ever built: arrange 118 elements by electron configuration and their chemical behaviour stops being a list.
An atomic number is a count of protons, and a neutral atom therefore has exactly that many electrons. Electrons fill shells in a fixed order, and the table is a picture of that filling order.
Read across a period and you are walking increasing atomic number through one shell being filled. Read down a group and you are walking elements that differ by one valence electron — electrons in the outermost shell. Group 1 members have one, group 17 have seven, group 18 have eight.
That single fact does most of the work in the table. Two elements in the same group have similar chemistry because they have similar valence configurations, not because of any deeper relationship between the nuclei.
The blocks are the shell structure made visible. The s-block is two groups wide because the first shell holds 2. The p-block is six wide because a p subshell holds 6. The d-block is ten wide because a d subshell holds 10, which is also why every d-block row is called a transition series. The f-block, printed below the main body as the lanthanides and actinides, is fourteen wide.
The two numbering systems answer different questions, and confusing them is the most common way to misread the table.
A period is a horizontal row, and its number tells you how many electron shells are occupied. Every element in period 3 has three occupied shells; period 4 adds a fourth. Period number is shell count.
A group is a vertical column, and its number tells you how many valence electrons an atom has — with the main-group exception that groups 13 to 18 are numbered one higher than the valence count. Group 17 is seven valence electrons, which is why the halogens need exactly one more to close the shell.
So group predicts chemistry, period predicts shell count and radius. Down a group you add a shell and the atom gets bigger with weaker outer electrons held less tightly; across a period radius shrinks because the rising nuclear charge pulls the same shell inward. That shrinkage is why lithium is reactive and fluorine is aggressive, despite both having a single valence electron.
Each main group is a family with a recognisable personality, and the personality follows from one electron short or one electron over.
Alkali metals (group 1) have a single loosely held valence electron and are the most reactive metals. They tarnish in air, react violently with water, and are too soft to have many uses in pure form.
Alkaline earth metals (group 2) have two valence electrons and are less reactive because removing both takes more energy. Magnesium sits in seawater.
Halogens (group 17) have seven valence electrons and need exactly one more, so they are the most reactive nonmetals and exist as diatomic molecules: F₂, Cl₂, Br₂, I₂. Fluorine is the most electronegative element in the table.
Noble gases (group 18) have a full valence shell, which is the whole reason for their inertness. Helium's 1s² is closed at two electrons — a full shell before any p subshell is used at all.
Transition metals (the d-block) have partially filled d subshells, which makes them variable-valent. Iron forms Fe²⁺ and Fe³⁺; that variable oxidation state is why these elements colour their compounds, catalyse reactions, and form the alloys that structural engineering depends on.
Mendeleev arranged the elements in 1869 by atomic weight and hit the same problem the table still frames: atomic weight does not increase cleanly across a row. Tellurium (127.6) sits heavier than iodine (126.9), and no amount of reordering by weight fixes the row without breaking something else.
He responded with two rules that a table without gaps could never have satisfied. Elements with similar properties must behave alike, and any gap must correspond to an undiscovered element whose properties can be predicted. The gaps he left — the region now occupied by the noble gases, plus scandium, gallium, germanium and a handful of others — turned out to be filled by elements close to his forecasts.
This is the strongest argument the table has. A classification that merely sorted known facts would be a convenience. One that predicted the atomic weight, density, melting point and oxide formula of an element before it was isolated was doing real chemistry.
Madelung's rule later gave the filling order a cause: orbitals fill by increasing n + ℓ, so 4s fills before 3d. That is why potassium and calcium come before the transition series in period 4, and why scandium starts the d-block with a nearly empty d shell and an electron that prefers to move into it. The usual 4s-versus-3d anomaly is a consequence of the rule plus energy lowering once the d subshell is occupied, not a failure of it.
Two boundaries of the modern table are practical rather than fundamental.
The heavy elements past bismuth are substantially relativistic. Inner electrons move fast enough for their mass to matter, which contracts s and p orbitals and expands d and f orbitals. Gold looks gold partly because relativistic effects shift its absorption; mercury stays liquid because its 6s shell is contracted and hard to bond with. Francium is predicted to be more reactive than caesium, and predicted to have a noticeably different appearance, for the same reason.
And the lanthanides and actinides are conventionally detached from the main body purely to keep the table rectangular. They belong in the gap between groups 2 and 3; the f-block exists so that group 3 can start a row that is ten columns wide.
The heavier synthetic elements also have names and masses that are provisional. Oganesson, for instance, is predicted to be a solid rather than a gas, on relativistic grounds that no measurement has yet confirmed.
The interactive table on this page encodes one number, one group, one period, one atomic mass, one phase at room temperature, and one category per element. Use it to read across, down, and diagonally.
Across a period, mass climbs while radius shrinks and metallic character falls. Down a group, radius grows and metallic character rises. Diagonally you trace similar sizes, which is why lithium and magnesium behave more like each other than either does to its own group — the diagonal relationship that gave germanium its use as a semiconductor substitute for silicon.