The fragrance pyramid is a diagram, not a formula. Nothing inside the bottle is stacked on three floors, and nothing waits its turn. Every molecule in a perfume begins evaporating the instant the spray lands, all of them at once, at rates set by vapour pressure and by how hard each molecule grips its neighbours. What you experience as top, heart and base is not a schedule anyone wrote. It is the residue of a race that everything entered simultaneously and that the light material lost first. What follows is the physics underneath the picture: where the three tiers came from, exactly where the diagram stops describing reality, and how to read a note list without being managed by it.
11 min
The Pyramid Was a Bench Tool Before It Was a Picture
Jean Carles, born 12 January 1892 and dead 8 July 1966, worked at Roure Bertrand Fils in Grasse and founded the Roure perfumery school, serving as its first director from 1946. He measured the relative volatility of raw materials systematically, then composed from the bottom up. Base first, then heart, then the fleeting top. He went anosmic late in life and kept composing from memory alone.
That order was a manufacturing discipline, not a description of perception. Build the slow material first, because it fixes a character you cannot revise later. Layer the volatile material on top, because it is cheap to adjust. The method described how a perfumer works at a bench. It never claimed to describe what a wearer smells at minute three.
The quantitative backbone came from W. A. Poucher, who ranked perfumery materials by evaporation coefficient on a scale of 1 to 100 — a scheme still reprinted in Poucher's Perfumes, Cosmetics and Soaps. Coefficients of 1 to 14 are classed top notes, 15 to 60 middle, 61 to 100 base. The cut points are administrative. Nothing in chemistry insists on three groups rather than five, or eleven.
Somewhere between the bench and the sales counter, the tiers became a triangle. A triangle is an excellent picture. It implies floors, sequence and hierarchy, and it prints cleanly on a display card. Bergamot goes at the apex, oakmoss at the foundation, and the customer receives a story with a beginning, a middle and an end. The teaching survived. The measurements it was built on did not travel with it.
The pyramid and the fragrance wheel are two different flattenings of the same object. Both leave out most of it.
Vapour Pressure, Not the Clock
Vapour pressure is the tendency of a liquid to escape into the air above it, measured at a given temperature. Higher vapour pressure, faster escape. That is the whole mechanism, and it does not require a perfumer's intent.
In a mixture, no component evaporates on its own schedule. Alírio Rodrigues, Idelfonso Nogueira and Rui Faria set it out plainly in Molecules in 2021: the gas-phase concentration of component i above a perfume is Cig = xiγiPisat/RT — mole fraction, times activity coefficient, times saturated vapour pressure, divided by RT. This is Raoult's law in its non-ideal form, and it governs every sprayed perfume on earth.
Read the equation for what it forbids. There is no time variable in it. Nothing in it permits a material to wait its turn. From the first instant, every component is present in the vapour above your skin, at a concentration proportional to how much of it is there and how badly it wants to leave. What changes over the following hours is xi, the mole fraction: the volatile species are leaving fastest, so they run out fastest.
Ethanol makes this concrete. At 46.07 g/mol it carries a vapour pressure of 59.3 mmHg at 25 °C. Limonene, the standard citrus terpene, sits at 1.55 mmHg. Ethanol is roughly thirty-eight times more eager to leave the liquid. The first two seconds of any sprayed perfume are dominated by solvent, not by fruit. Every wearer has smelled this, and almost none have been told what it was. The flash the trade calls "the opening" is, chemically, mostly the thing added to keep the composition liquid.
Molecular Weight Is a Proxy, and a Bad One
Molecular weight tracks volatility because heavier molecules present more surface for van der Waals contact, so more energy is needed to pull one free. The trend is real. Almost everything you can smell weighs under about 300 g/mol — above that, too little of the substance reaches the receptor sheet in your nose to register at all. Inside that narrow band, though, weight predicts badly.
| Material | Molecular weight (g/mol) | Vapour pressure at 25 °C (mmHg) | Usual pyramid tier |
|---|---|---|---|
| Ethanol | 46.07 | 59.3 | never listed |
| Limonene | 136.23 | 1.55 | top |
| Linalool | 154.25 | 0.159 (at 23.5 °C) | top / heart |
| Vanillin | 152.15 | 1.18 × 10⁴ | base |
| Galaxolide | 258.4 | 5.4 × 10⁴ | base |
Vanillin weighs 152.15 g/mol. Linalool weighs 154.25. Two grams per mole apart. Linalool's vapour pressure is 0.159 mmHg, measured at 23.5 °C by Li and colleagues in Environment International in 1998. Vanillin's is 1.18 × 10⁴ mmHg at 25 °C, from Yaws's Handbook of Vapor Pressure. A factor of roughly 1,350 between two molecules of practically identical mass.
Now run it backwards. Galaxolide weighs 258.4 g/mol, seventy per cent heavier than vanillin, and its vapour pressure is 5.4 × 10⁴ mmHg, reported by Balk and Ford in Toxicology Letters in 1999. The heavier molecule is about four and a half times more volatile than the lighter one.
The reason is not mass. It is hydrogen bonding. Vanillin carries a phenolic hydroxyl and an aldehyde, and its molecules grip one another so firmly that vanillin is a crystalline solid at room temperature, melting at 81.5 °C. Galaxolide is a greasy polycyclic ether with almost nothing to grip with. Ask what a molecule weighs and you get a hint. Ask what it holds onto and you get the answer.
Concentration is the other number people misread on a box. An eau de toilette can outlast an extrait, and often does.
Your Base Note Arrived at Second One
| What the diagram implies | What the physics gives |
|---|---|
| Top notes play, then stop | Every component evaporates from second one; the volatile ones simply exhaust their supply first |
| Base notes appear after several hours | Base notes were in the vapour immediately, below the perceptual threshold set by louder material |
| Three discrete phases | One continuous drift in mole fractions, with no boundaries anywhere in the curve |
| The composition evolves | The composition depletes; what changes is which part of it you can still hear |
A patchouli base note is not backstage during hour one. It is on your arm, evaporating, entering the air above your wrist at a low but strictly non-zero rate from the moment the spray lands. You do not notice it because limonene and benzyl acetate are shouting over it at concentrations thousands of times higher.
The pyramid is not a relay race. It is an orchestra in which the piccolos run out of breath. The cellos were playing the whole time.
This has a consequence most people get backwards. The drydown is not a phase the perfumer scheduled for hour four. It is what remains audible once the loud material has left the building. When a vetiver or a sandalwood seems to "emerge" at hour three, nothing emerged. Your access to it changed.
Rodrigues and colleagues give the conventional durations: 15 to 30 minutes on skin for top notes, 3 to 4 hours for middle, beyond 4 hours for base. Treat those as averages and they are useful. Treat them as promises and they will fail you, because vapour pressure climbs steeply with temperature. The same formula reaches its drydown faster on a wrist in July than on a card in a shop in January.
Volatility and Potency Are Separate Axes
How much of a molecule reaches your nose is one question. How little of it you need in order to notice is an entirely different one, and the pyramid collapses the two.
The industry measure of potency is the odour detection threshold, the gas-phase concentration at which a material becomes perceptible. Rodrigues and colleagues list limonene at 6.19 × 10⁻¹ mg/m³ and linalool at 9.33 × 10⁻³ mg/m³. Linalool is roughly sixty-six times more potent by threshold. Two molecules in the same weight class, both routinely printed as top notes, and one registers at a sixty-sixth of the concentration the other needs.
The quantity that actually predicts perception is the ratio of the two: Odour Value, defined as gas-phase concentration divided by detection threshold. Perceived intensity then follows Stevens's power law, ψi = (Cig/ODTi)n, with a fractional exponent — doubling the dose of a material does not double its loudness. Overdosing is a poor way to buy presence.
This is how a nominal base note dominates an opening. A material with a low vapour pressure and a brutally low threshold can carry a higher odour value at minute one than a citrus present at ten times the dose. Indole does precisely this in white florals: heavy, slow, and perceptible where most materials are silent. A jasmine accord announces itself immediately, and the pyramid files it under heart.
At Première Peau we built Gravitas Capitale on exactly this asymmetry — cedrat, the thick-rinded citron, set against asphalt. On a diagram, one is a top note and the other has no tier at all. On skin they arrive together, because the mineral register is working at an odour value the citrus cannot drown.
Some flowers give up no oil at all, so their scent is captured as data instead. Headspace analysis reads a living flower without touching it.
Three Tests for Any Note List
Test one: check whether the material can be extracted at all. Lily of the valley has no commercial essential oil. Distillation, solvent extraction and supercritical CO₂ all fail to recover a usable quantity, and CO₂ destroys the balance that makes the flower recognisable, as Perfumer & Flavorist has documented. Every muguet note ever sold is a reconstruction, assembled from headspace analysis of the living plant. The same holds for gardenia, lilac and freesia. A note list printing them is naming a target, not an ingredient.
Test two: read for the omissions. Hedione, the jasmine-facet ester Edouard Demole synthesised at Firmenich in 1959, weighs 226.31 g/mol. The University of Bristol's Molecule of the Month entry records documented use levels in commercial fragrances running from 2.5 per cent up to 65 per cent of the concentrate. A material can be two-thirds of a formula and never appear on the pyramid once. The same goes for Iso E Super at 234.38 g/mol, ambroxan at 236.39 and ambrettolide at 252.39. These are the load-bearing walls. The note list describes the wallpaper.
Note lists are generally written after the formula is finished, by people whose job is the story rather than the chemistry. That is not deception. A manifest of thirty-five chemical names would tell a shopper nothing usable, and an evocation is a legitimate translation of a smell. But translations lose things, and this one loses the structure. Read a note list as a mood board, never as an inventory.
Test three: check the tier against the material. If a list places a musk or coumarin among the top notes, either the copy was improvised or the material was chosen for a reason the diagram has no vocabulary for. Both are worth knowing before you buy.
Vapour pressure decides what leaves your skin; air currents decide where it goes. Sillage is the second half of the same physics.
Run It on Your Own Arm
The claim in this article is falsifiable in a single day, using one bottle and one strip of paper.
Spray once on the inside of your wrist and once on a blotter. Smell both at thirty seconds, two minutes, ten minutes, thirty minutes, two hours and six hours. Do not compare wrist to blotter. Compare each one to itself an hour earlier, which is the only comparison your nose can make honestly.
Two things will happen. The blotter will hold the light material noticeably longer than your wrist, because paper is cooler and does not perspire. And the clean three-act sequence you were promised will not arrive. You will find the iris or the leather already faintly legible at minute two, under the citrus, if you look for it rather than waiting for it.
Then run the real test. At hour four, when the volatile material is gone and only the slow molecules remain, ask whether what you are smelling was on the printed list at all. That gap — between the triangle on the card and the orris root, the rose or the white truffle still working on your wrist — is the whole subject of this article. The diagram is a map drawn for a customer. The evaporation curve is the territory, and it starts at second one.
A low threshold beats a high dose every time. Insuline Safrine is built on saffron, a material that needs almost none of itself to be heard.
Frequently Asked Questions
What are scent notes in perfume?
Scent notes are named descriptors for the smells a perfume presents over time, grouped conventionally into top, heart and base. They are perceptual labels, not a list of the materials in the bottle. A single printed note such as saffron may be one molecule, a natural extract containing hundreds, or an accord built to evoke it.
Is the fragrance pyramid real?
The volatility differences it describes are real and measurable. The three-floor structure is not. Materials do not take turns; they all evaporate from the first second, and the tiers are administrative cut points on a continuous scale of vapour pressure that W. A. Poucher divided at 14 and 60 on a scale of 100.
What decides whether a note is top, heart or base?
Vapour pressure at skin temperature, moderated by how strongly the molecules attract one another. Hydrogen bonding matters more than mass: vanillin and linalool differ by two grams per mole and by a factor of roughly 1,350 in vapour pressure, because vanillin's hydroxyl and aldehyde groups lock its molecules into a crystal that melts at 81.5 °C.
Do base notes really only appear after several hours?
No. A base note is in the air above your skin from the moment you spray, at a concentration set by Raoult's law. It becomes noticeable later only because the volatile material that was masking it has evaporated away. Nothing emerged; the noise stopped.
Does molecular weight determine how long a perfume lasts?
Only loosely, and it fails at the edges. Galaxolide weighs seventy per cent more than vanillin yet is about four and a half times more volatile. Weight sets a rough ceiling — almost nothing above 300 g/mol has any smell at all — but within the odorous range, intermolecular forces dominate.
Why do two perfumes with identical note lists smell different?
Because the list names impressions, not quantities or materials. Two compositions can both print "jasmine, cedar, musk" while sharing no molecules, and the proportions, which are never disclosed, do most of the work. Materials such as Hedione or Iso E Super can occupy a large share of a formula without ever being named.
Why are notes listed that cannot be extracted from the plant?
Some flowers yield no usable oil by any method. Lily of the valley, gardenia, lilac and freesia are reconstructions, built by analysing the vapour above the living flower. Printing the flower's name is shorthand for the intended effect, not a claim about the contents.
How long do top, heart and base notes actually last on skin?
The conventional figures published by Rodrigues and colleagues are 15 to 30 minutes for top notes, 3 to 4 hours for heart notes and beyond 4 hours for base notes. Those are averages across formulas. Your own result will shift with skin temperature, ambient temperature and how much you applied, because vapour pressure rises steeply with heat.
