The property the data almost always indicates is cohesion, measured as surface tension — and the molecular explanation is hydrogen bonding. If a table shows water's surface tension at roughly 72 mN/m at 25 °C next to acetone's 23 or ethanol's 22, and notes anomalies like an unusually high boiling point or a high specific heat, the pattern the data is pointing at is strong intermolecular attraction pulling water molecules into each other. Cohesion is that attraction; surface tension is its measurable expression at an interface. The rest of this article shows how to prove that link, what the standard numbers are, and where the neat story breaks down.

The Core Answer: Which Property the Data Describes

When a dataset on water reports a high boiling point, a high specific heat capacity, and a high surface tension side by side, the shared variable is cohesion — the tendency of water molecules to stick together. Cohesion arises because each H₂O molecule carries two partially positive hydrogen atoms and one partially negative oxygen atom, so neighbouring molecules attract in a persistent network of hydrogen bonds. That attraction is what the data is really capturing whenever it flags water as "abnormally" sticky or "abnormally" stable.

Untrained readers often read surface tension as a skin on top of the water. It isn't. The surface tension value describes how much extra energy is needed to create one more square metre of liquid surface, because molecules at the surface lose some of the bonding they would enjoy in the bulk — that energetic cost is what the force measurement reflects.

Running a quick comparison is the cleanest way to confirm the interpretation. Ethanol has a similar molecular mass but only one hydrogen donor, so water's surface tension sits around three times higher at room temperature. If the property in your data leaves water with a significantly higher boiling point and heat capacity than molecules of comparable size, cohesion is the coherent label.

Match the property to its measurement before drawing conclusions. Boiling point, specific heat, surface tension, viscosity, and the enthalpy of vaporisation all rise because of hydrogen bonding, but each responds differently to temperature and to dissolved substances, so a dataset that includes plain surface tension alone cannot prove the rest.

The Standard Numbers: A Reference Table for Water

Comparative figures make the cohesion claim testable. Since no current source can be fetched here, the table below uses textbook-standard values at 25 °C from physical chemistry handbooks such as the CRC Handbook of Chemistry and Physics; treat those constants as stable, and verify any figure your lab will publish against the current edition.

Substance (at 25 °C unless noted) Surface tension (mN/m) Boiling point (°C) Specific heat capacity (J/g·°C)
Water 72.0 100 4.18
Ethanol 22.1 78.4 2.44
Acetone 23.3 56.1 2.15
Mercury 485.5 356.7 0.140
Hexane 18.4 68.7 2.26

To test the correctness of the surface tension value from your own data, drop a temperature at which the measurement was taken. Water's surface tension falls roughly 0.15–0.16 mN/m per °C near room temperature (CRC's tabulated series show about 71.99 mN/m at 25 °C and 71.18 mN/m at 30 °C), so an observed 69.5 mN/m is reasonable at around 35–38 °C and is not evidence of contamination. Seven or eight mN/m low at a stated 25 °C would be.

Mercury complicates the simple ranking because its cohesion comes from metallic bonding rather than hydrogen bonding; text it out of any rank without mentioning the mechanism, or a reader will reasonably ask why the "most cohesive" liquid on the table has no hydrogen atoms at all. Being explicit prevents the table from implying that high surface tension always means hydrogen bonding.

Cross-checking the two correlates above states the property plainly: water's boiling point is 21.6 °C higher than ethanol's even though ethanol's molar mass of 46.07 g/mol is larger than water's 18.02 g/mol. The size argument predicts the wrong order, cohesion predicts the right one, and that reversal is the single most useful line for a student defending the interpretation.

Explaining the Property From Molecular Structure

The explanation starts with geometry. Each oxygen atom holds two pairs of non-bonding (lone) electrons, and each O–H bond is polarised toward oxygen — because oxygen is markedly more electronegative than hydrogen — giving the molecule two partially positive hydrogen ends and two partially negative lone-pair ends. One water molecule can therefore accept two hydrogen bonds through its lone pairs and donate two through its hydrogens.

Hydrogen bonds are weak individually but cooperative in bulk. Each bond has an energy on the order of 20 kJ/mol in water, over ten times weaker than a typical O–H covalent bond, but each molecule is shifted into about four such links at a time — two donor, two acceptor — in liquid water, so the average number of hydrogen bonds per molecule is roughly 3.4–3.5. Short-lived bonds breaking and reforming on the scale of picoseconds create a dense, constantly reorganising network rather than a fixed lattice.

That is why the bulk properties come out high. Pulling a water molecule out of the liquid and into vapour — the enthalpy of vaporisation, 40.7 kJ/mol at 100 °C — means breaking the molecule's share of that network. Heating the liquid means paying energy to excite those same intermolecular links before the molecules can move faster, so water's specific heat stays near 4.18 J/g·°C, and boiling requires disordering them at 100 °C.

The structure story only applies to the liquid because the liquid is a network. In ice the same molecules build an ordered tetrahedral lattice with near-complete hydrogen bonding, which is why ice is less dense than liquid water: the ordered framework holds molecules further apart than the maximally packed liquid arrangement at 0 °C. Any dataset that shows ice floating alongside high cohesion and high heat capacity fits the same molecular picture.

How to Prove the Mechanism in the Laboratory

A single property value can be explained after the fact by almost any story. The convincing projects hold two variables at once, which is why classroom cohesion experiments pair a contact-angle or drop test with a detergent addition or a temperature ramp.

1: Measure surface tension directly, using a ring or Wilhelmy-plate tensiometer if one is available, or photograph a calibrated drop and apply the Young–Laplace relation for a fallback figure. 2: Repeat the measurement across a temperature range of about 10–40 °C, recording each pair, and plot the values. 3: Confirm the slope: expect a decrease of roughly 0.15–0.16 mN/m per °C, and note whether a phase change or evaporation distorts the top of the range. 4: Add a fixed drop of detergent to a fresh sample and remeasure to show surfactant molecules sitting at the interface and lowering the value. 5: Repeat a cold control drop of clean water beside a drop of the detergent solution on the same plastic tray.

If the values scatter wildly from run to run, a dirty ring or plate is the usual culprit, and wiping it with solvent between readings often fixes the drift. When the drop test on a plastic tray gives inconsistent results, the clue to contamination is usually a drop that beads on part of the surface and spreads on another; clean both surfaces identically before repeating.

Log both the temperature and the surfactant mass fraction next to every reading; a bare force value is untraceable later. The pattern the experiment is meant to demonstrate — temperature sensitivity and surfactant sensitivity — is what connects the raw number to the hydrogen-bond explanation.

Where the Simple Story Breaks: Temperature Sensitivity and the Air Interface

The textbook picture of a stabilising hydrogen-bond network is right about the average, wrong about the detail. The average lifetime of a single water hydrogen bond is on the order of a picosecond in liquid water, so the network is not a permanent scaffold. The best-supported structural model is a fluctuating, distorted three-dimensional network with a broad distribution of local environments rather than uniformly four-bonded molecules.

The surface is a distinct environment, and datasets that pool "the" surface tension of water are really reporting a property of the liquid–air interface. Adsorbed contaminants from the air reduce the value, which is why careful measurements use freshly drawn high-purity water, a clean platinum ring or plate, and immediate readings. A single table row that prints two significant figures without a measurement method tells you less than it appears to.

Temperature exposes a subtler wrinkle. The enthalpy of vaporisation falls from roughly 40.7 kJ/mol at 100 °C to about 43.99 kJ/mol near 25 °C — the same value quoted at two different reference temperatures describes two different balances of bonds. Conflicting published figures on water's cohesion usually trace back to reference temperature and measurement method rather than disagreement about the underlying chemistry.

At the high-temperature end, cohesion weakens enough that water behaves increasingly like a simple liquid, and the "anomalous" heat capacity drops. If your dataset runs past about 250 °C, the explanation shifts from "hydrogen-bonding-dominated" to "network breaking up", so don't apply the room-temperature breakdown to a superheated value without checking the state.

The Densities and Heat Capacities That Break the Simple Correlation

Water's most famous anomaly is the density maximum at about 4 °C, where the density reaches roughly 0.99997 g/cm³, slightly below the value near 4 °C quoted in some tables as 1.000 g/cm³ at 4 °C; the maximum is at the latter and the linear term is small. That maximum is a direct consequence of two competing effects: the near-tetrahedral hydrogen-bond framework tends to hold molecules apart, while ordinary thermal contraction tends to draw them together.

The heat capacity of water is large enough to matter on a household scale. It takes about 4.18 joules to raise one gram of water by 1 °C, while heating the same mass of ethanol needs roughly 2.44 J and acetone needs roughly 2.15 J. That is why a warm-water bottle holds heat longer than a warm oil bottle of equal mass, and why a home water-heating device adds energy into the water's hydrogen-bonded network.

Heat capacity also gives you a cheap way to check a dataset's internal consistency. If a computed change in heat capacity for a water sample matches the expected 4.18 J/g·°C within rounding, the other properties listed for the same sample are more likely measured under the same conditions. If the specific heat is far off the standard line, the measurement or the sample is suspect.

Ice density fits the same molecular picture. The ordered tetrahedral lattice in ice has fewer molecules per cubic centimetre than the room-temperature liquid, which is why ice floats rather than sinking; a dataset that shows the low-density solid alongside the high surface tension records two expressions of the same ordered hydrogen-bonding structure, one in the solid and one at the liquid surface.

How Dissolved Substances and Surfactants Interfere With the Measurement

Additives can move the measured property substantially without changing the structure story. When soap molecules occupy the interface, water's surface tension drops from roughly 72 mN/m to values in the low twenties — the same range the simple liquids above occupy — so a surface tension measurement on an unknown sample is only identification-grade if the sample is prepared clean.

Biological samples are the most common case. Protein solutions and cell culture media read low in surface tension because proteins and lipids accumulate at the air–water interface. A tabulated "surface tension of water" printed next to a serum experiment is therefore misleading unless the sample is clarified.

Dissolved salts behave differently. A simple salt like sodium chloride slightly raises surface tension because ions pull water molecules more tightly into the bulk and are excluded from the interface. This is the opposite direction to detergent, and that contrast is one of the clearest signs that the number is measuring the interface and the hydrogen-bond network at once.

Insight from the sample context matters here. A clean, freshly prepared buffer reads near the standard line, while a buffered solution with an added detergent or a protein reads low, and the difference separates a solvent effect from a surfactant effect; state which one you are reporting or the reading becomes meaningless.

Putting the Numbers Together: A Worked Arithmetic Check

The strongest argument for cohesion as the property being described is arithmetic that the alternative explanations cannot make. Since water's molar mass (18.02 g/mol) is less than half of ethanol's (46.07 g/mol), a molecular-size argument predicts water should boil lower than ethanol by a measurable margin. Instead, water boils 21.6 °C higher at 1 atm, so the size argument is the wrong model.

The molecular explanation supplies what the size argument lacks. Using water's standard specific heat of 4.18 J/g·°C, heating 250 g of water from 20 °C to 100 °C requires about 83.6 kJ of energy; heating the same 250 g of ethanol through the same 80 °C rise needs about 48.8 kJ, and acetone about 43.0 kJ. The 34.7 kJ difference between water and ethanol is a real cost in the kitchen or in a lab, not a puzzle, because each degree of water's warming is paid for by the hydrogen-bond network.

Derive a second quantity from the table rather than quoting it. The heat needed to warm that same 250 g of water for another 80 °C is smaller than the total heat of vaporisation of the same mass at 100 °C — the vaporisation energy (40.7 kJ/mol times 13.9 mol for 250 g is roughly 564 kJ) is far larger than the 83.6 kJ of sensible heating, which is why steam burns are severe and why boiling water takes longer than simply heating it. That same excess pays for breaking the liquid's hydrogen bonds, which connects vaporisation directly to the cohesion explanation.

Check a second dimension. The specific-heat ratio between water and ethanol at room temperature is 4.18 / 2.44 ≈ 1.7, and the enthalpy-of-vaporisation ratio is 40.7 / 38.6 ≈ 1.05 (ethanol's value near 40 °C is about 38.6 kJ/mol, so the two liquids are closer in vapour energy than in heat capacity). The two properties don't move in lockstep, which is why a dataset that reports only one of them supports a weaker claim than one that reports both.

FAQ

What is the most direct measurement of water's cohesion?

Surface tension, because it reports the energy cost of creating new liquid surface and that cost comes from molecules at the surface having fewer hydrogen bonds. A clean value at 25 °C is about 72.0 mN/m. Boiling point and heat capacity are indirect reflections; they are explained by the same bonds, but they depend on how much of the network a given process disturbs.

Does cohesion explain why ice floats?

Not by itself. Ice floats because the ordered tetrahedral hydrogen-bond lattice in ice packs molecules less densely than the disordered liquid at 0 °C. The molecular cohesion that produces high surface tension also drives the ordered lattice, so the two facts share a cause but aren't the same property. A dataset showing ice density less than liquid water is evidence of the lattice, not evidence of cohesion.

Why does detergent lower the surface tension so sharply?

Detergent molecules (surfactants) migrate to the air–liquid interface and replace water–water hydrogen bonds with weaker water–surfactant contacts. Measurements can fall from about 72 mN/m toward the low twenties. If your sample has been exposed to soap, protein, or a cleaning solution, re-measure with fresh water before reporting a value.

Is water's surface tension unusually high compared with molecules of the same size?

Yes. Ethanol and acetone are close to water's molar size, yet their surface tensions sit near 22–23 mN/m against water's 72.0 mN/m at 25 °C — a factor of about three. The comparison is only fair when molecular mass, temperature, and purity are matched, since both the reference temperature and the presence of surfactants shift the result.

Can water's cohesion be measured at home?

Yes, roughly, with a drop-shape or contact-angle test on a calibrated surface. The catch is calibration: a phone photo of a bead of water on a plastic plate gives a plausible but uncalibrated contact angle that cannot be reported as a surface tension value. Do not publish a home measurement as an absolute surface tension; use a tensiometer or the drop-shape method with known standards.

How does hydrogen bonding produce surface tension at the air interface?

A water molecule at the surface can hydrogen-bond with neighbours only on one side, so it has a lower average bond count than a molecule in the bulk. Surface energy is the fraction of a bulk molecule's bonds that a surface molecule loses. Because each water molecule can donate and accept about two bonds each, the loss per surface molecule is large — and that is what makes the measured number high.

Do superheated or very hot water samples follow the same explanation?

Only partly. Near and above about 250 °C the hydrogen-bond network weakens and water's properties shift toward a simple liquid, with heat capacity falling below the room-temperature value. A superheated reading should not be explained with the room-temperature network without checking the stated temperature and pressure; the correct check is the measurement conditions, not the mechanism.

Does cohesion have a role in the water's high specific heat, or are those separate properties?

They share a cause but are not the same property. Specific heat measures how much energy is needed to raise a mass by one degree and reflects the energy needed to excite the hydrogen-bond network; surface tension measures energy per unit of new surface. A change in temperature moves both, but not by the same amount, so a dataset needs both properties to support a strong cohesion claim.

The discussion above gives general scientific explanation only; for a laboratory result to be reported formally, confirm every constant against your measurement conditions and a current, citable physical chemistry reference.