Melting point generally decreases down a group for metals, but increases sharply for nonmetals and metalloids. There is no single monotonic rule because melting is governed by which force holds the solid together: metallic bonding, covalent network bonding, or intermolecular forces. The trend's direction depends on which of these dominates the group in question.

The mechanism: melting reflects the force holding the solid together

Melting happens when thermal energy defeats the ordered lattice. The temperature at which this occurs is set by the strength of whatever bonds or forces hold the particles in place. That makes melting point a reporter of bonding type, not an intrinsic property of an element's position on the table. Chemists call this the character of the solid state.

Three bonding regimes produce three different down-group behaviors. In metallic solids, atoms release valence electrons into a shared sea, and the lattice energy scales with how tightly the positive cores hold that electron sea. In covalent network solids, every atom is locked into a continuous framework of directional bonds, so melting requires breaking the network itself. In molecular solids, discrete molecules are held to each other by comparatively weak dispersion and dipole forces, and the network of molecules, not the molecule, breaks on melting.

Each regime responds differently as you move down a group, because atomic size, nuclear charge, and electron screening all change together. The dominant variable is rarely the same across the periodic table. One plain sentence: melting point tells you what kind of glue holds the solid together, and that glue changes as you descend.

Metallic groups show the clearest decrease. In group 1, lithium melts near 180 °C (454 K) and cesium near 28 °C (301 K). In group 2, beryllium melts near 1,287 °C and barium near 727 °C. The pattern is not perfectly smooth: magnesium melts higher than calcium, which is why the trend is described as general rather than exact. Across the metallic groups overall, melting points fall from hundreds or thousands of degrees at the top to tens or low hundreds at the bottom.

Covalent network groups run the other way. Carbon's diamond form is a giant covalent lattice that does not melt under ordinary conditions; it sublimes near 3,600 °C. Silicon melts near 1,414 °C and germanium near 938 °C, so the melting points of the network-forming elements in that column are actually quite high. This is a useful counterpoint: carbon does not fit neatly into the "melting point decreases" story because its solid is held by covalent bonds, not metallic ones. The group 14 trend is also complicated by the fact that the lighter elements are network solids while the heavier ones become more metallic and less refractory.

Molecular groups are the sharpest case. Halogen melting points rise down the group: fluorine melts near −220 °C, chlorine near −101 °C, bromine near −7 °C, and iodine near 114 °C. Noble gases show the same direction: helium melts only near absolute zero, while radon melts near −71 °C. In these groups the trend is upward, because larger atoms are more polarizable, and dispersion forces between molecules strengthen with size.

Group type Example group Down-group direction Dominant force Typical range
Alkali metals Group 1 Decreases Metallic bonding 28–180 °C
Alkaline earths Group 2 Decreases Metallic bonding 727–1,287 °C
Halogens Group 17 Increases Dispersion forces −220 to 114 °C
Noble gases Group 18 Increases Dispersion forces Near absolute zero to −71 °C
Network solids (partial) Group 14 Mixed; generally very high Covalent network 938–3,600 °C

The numbers show that the keyword question has no universal single answer. Direction depends on bonding class, and the boundary between classes is exactly where the trend can flip or flatten.

Edge cases and the counterintuitive detail

The counterintuitive fact is that a larger atom can produce a higher melting point. In molecular solids, bigger, more polarizable atoms generate stronger London dispersion forces, so iodine melts far above fluorine despite having weaker intramolecular bonds and a larger radius. Size alone does not predict melting; the relevant quantity is how strongly the solid's units attract each other, not how strongly any single bond holds a molecule together.

Several other edge cases deserve attention. Gallium melts near 30 °C, so it liquefies in a warm hand, and it is adjacent to elements whose melting points differ by hundreds of degrees. Mercury is liquid at room temperature, while its group neighbor copernicium is predicted to behave differently; predictions for superheavy elements rest on relativistic effects and are not measured as reliably as those for lighter elements. Carbon's diamond and graphite forms melt at dramatically different conditions because the solid structures differ, not the element.

Another professional caveat: melting point is sensitive to pressure and to impurity, which is why reported values for the same element can differ between sources. Many tabulated values are at standard pressure, and some are extrapolated. For the heavier elements, especially the synthetic ones, values are predictions rather than measurements, and you should treat those numbers as uncertain.

Misconceptions and a checklist of red flags

A common misconception is that melting point must decrease down every group because atomic radius increases. That logic only applies where the lattice is held by forces that weaken with distance, such as simple ionic and many metallic systems. It fails for molecular solids, where larger units interact more strongly.

A second misconception is that a high melting point means strong chemical bonds inside the molecule. In molecular solids, the bonds inside each molecule can be strong while the forces between molecules are weak. Melting tests the between-molecule forces, not the within-molecule bonds. This is why ice melts near 0 °C while the O–H bonds in water remain intact.

Red-flag warnings when reading melting-point claims:

A source states a single "group trend" without naming the bonding class, which hides the molecular-versus-metallic split. A table gives values for superheavy or synthetic elements without marking them as predicted or extrapolated. Melting points are quoted without pressure conditions or purity notes, making comparison unreliable. A review treats carbon, silicon, and germanium as a single clean trend when their solids differ in kind. Numbers are presented as current measurements when they originate from older compilations.

One plain sentence: if a claim about melting point does not say what kind of solid it is talking about, it is probably too broad.

Practical actions: using the trend correctly

To answer "does melting point increase or decrease down a group?" for a specific group, identify the solid type first. For any element, classify it as metallic, covalent network, or molecular, then apply the matching rule. For most columns of metals, expect a decrease; for halogen and noble-gas columns, expect an increase; for group 14 and similar mixed columns, expect no single direction and check each element individually.

Four concrete steps make this reliable:

Classify the solid type of the element, and of its lighter group neighbors, using a standard reference table. For metals, compare the elements across one period as well, because periodic trends in melting point are often more informative horizontally than vertically. For molecular solids, rank by molecular size and polarizability, not by radius alone. Treat any value for a synthetic element as provisional unless a measurement is reported with conditions.

Thresholds worth remembering: the metallic-decrease pattern is most trustworthy for groups 1 and 2, where the lightest element melts above 180 °C and the heaviest alkali metal melts below 30 °C. The molecular-increase pattern is clearest for group 17 and group 18. Group 14 is the key exception zone, with melting points spanning roughly 938 °C to about 3,600 °C depending on structure.

Limitations and what cannot be concluded

Melting point is a bulk property shaped by lattice energy, packing, and defects, so it cannot be predicted from periodic position alone. Databases of melting points are compiled at different pressures and purities, and some entries are calculated rather than measured. Superheavy elements have no reliable macroscopic samples, so their melting points are extrapolations based on relativistic quantum chemistry, not observations.

A single sentence cannot capture a trend that is not monotonic, and that is the honest answer to the keyword question. The decrease down a group is real and well measured for the metallic columns; the increase is real and well measured for the molecular columns; the boundary cases such as group 14 resist both. For any specific element, consult a primary reference and note the pressure, purity, and whether the value is measured or predicted. General periodic tables are useful for orientation, but they should not be treated as a substitute for element-specific data when accuracy matters.