A repeatable method for reasoning about a bottle nobody has explained to you — and three oils it is run on, out loud, from first principles.
~13 min read
Someone hands you a small amber bottle with a word on it you have never seen. There is no lesson for it, nobody in the room knows anything about it, and you are the one expected to have an opinion. The next two minutes are the point of this course.
Several thousand plants have been distilled for their aromatic fraction; a few hundred are traded in quantity. This app holds a few dozen. That gap will never close, which is why a method beats a list.
A list answers only the questions somebody anticipated. A method answers the ones nobody did — and you already hold every piece of one: what a molecule is, the carbon skeletons, the endings hung on them, and the families that sort those endings into behaviour.
By the end of this you should be able to take an unfamiliar bottle and say, before opening it, how fast it will evaporate, roughly what it will smell like, how you would dilute it, and what its families make you cautious about.
One boundary first, because it is the part people get wrong. The method does not tell you what an oil is for. Chemistry predicts how a substance behaves — how volatile, how reactive, how gentle, how persistent. Purpose is a different question.
The order is not decoration. Each step narrows what the next can be, so running them out of sequence means guessing at something you could already have known.
What you end with is a behaviour profile: a speed, a smell, a dilution, a set of cautions, a shelf position, a few partners. Not a purpose.
Diagram to come
The seven steps as one descending sequence down the page, each a numbered band with its question on the left and what it yields on the right, and a narrowing shape — bands getting visually tighter as you descend — so the eye reads a funnel: broad botanical facts at the top, a specific behaviour profile at the bottom. Below the last band, set apart by a rule, the output pair: what you end with, and the explicit limit. That final pair is the most important part of the drawing and must not shrink to a footnote. Drawn in code so every word can be translated.
Knowing the seven steps is not the same as running them. Each has a characteristic mistake.
| Dominant family | How it tends to behave | Dilution instinct |
|---|---|---|
| Monoterpenes and monoterpenols | Fast to moderate; monoterpenes oxidise with age | Standard; buy fresh |
| Esters | Sweet, usually mild — with sharp exceptions | Standard, unless the acid half is aromatic |
| Aldehydes and ketones | Intense from very little; care with young children | Heavier than you expect |
| Oxides | Sharp, opening; care near a child’s face | Standard |
| Phenols and phenylpropanoids | Hot on skin; the most potent group | Heavy — 1 drop in 10 or more |
| Sesquiterpenes and sesquiterpenols | Slow, deep, persistent | Standard; ages well |
What has the bottle actually told you?
How far you get depends on how much the label gives up.
Sometimes the botanical name is right, the family is right, and the oil is still not what you predicted. The usual reason is a chemotype: one species producing materially different oils depending on where and how it grows.
Thymus vulgaris grown hot and low yields an oil dominated by thymol, a phenol — hot on skin, heavily diluted. Grown high and cool, the same species can yield one dominated by linalool, a monoterpenol, gentle by comparison. Rosmarinus officinalis gives cineole, camphor and verbenone types; this library’s basil is the linalool type, at 40 to 80 per cent.
For most oils species is enough. For these few it is not, and the chemotype — ct. followed by the dominant compound — is the field that predicts behaviour. Where none is stated, assume the stricter possibility.
A 15 mL bottle, pale green-brown, labelled Geranium. Run the seven steps before reading a word of its library entry.
Step one. Pelargonium graveolens, family Geraniaceae — and the first step has already caught something. This is not a geranium. True geraniums are cranesbills, in the genus Geranium; this is a pelargonium, mislabelled by European gardeners three centuries ago. Geraniaceae is small and not one of the great aromatic families, so it gives less help than Lamiaceae.
Step two: flower and leaf, which rules out a base note and cold pressing. Step three confirms it: steam distilled, so no waxes, no pigments, no furocoumarins, no photosensitivity question. Two steps in, before any chemistry, the oil is a middle note with no sunlight caution.
Step four. The dominant family is monoterpene alcohols — citronellol, geraniol and linalool, around 154 to 156 grams per mole. Predict moderate volatility, a middle note lasting an hour or two, a sweet rose-like aroma, and a family among the gentler ones on skin.
Step five earns its keep. No second family is listed, so the reasoning comes from the compounds — and geraniol and citronellol are both among the fragrance compounds published cosmetic labelling rules require to be declared by name above a threshold, as recognised contact sensitisers. So soften step four: a monoterpenol oil that still wants dilution. Step six: dilute, no sunlight caution, possible skin sensitivity. Step seven: middle note, ordinary shelf life.
Now check it. The library gives Pelargonium graveolens, Geraniaceae, flower and leaf, steam distilled, middle note, monoterpene alcohols, aromatic and topical only, and a caution reading possible skin sensitivity — the prediction holds on every field. Shared linalool predicts lavender and clary sage; shared citronellol and geraniol predict rose. Four of the five listed partners come out of the profile; the fifth is frankincense, from step seven — a base note anchoring a middle note.
Where it missed. No ranges are published for geranium here, so the balance is unknowable — and whether citronellol leads or geraniol does is the difference between a green-rosy oil and a sweet-rosy one. It missed origin too: Réunion, Egyptian and Chinese geranium differ enough in that ratio that the trade prices them separately.
A 15 mL bottle, thick and very dark, labelled Vetiver. Same seven steps — and this time the first fails on purpose.
Step one: Vetiveria zizanioides, family Poaceae — the grasses, the same family as lemongrass and citronella, both bright, fast oils led by aldehydes and small alcohols. Predict from family alone and you are badly wrong.
Step two corrects it in one word: root. Vetiver is a grass whose roots go down two to three metres, and what a plant stores that far underground is not what it puts in a leaf. Roots, woods and resins hold large, low-volatility molecules, so every prediction moves — base note, viscous, dark, low yield, slow.
Step three: steam distilled, but a root distillation is not a leaf distillation. The roots are chopped and soaked first, and the run takes twelve to twenty-four hours where lavender takes under one. Step four confirms it: sesquiterpenols and sesquiterpenes — fifteen carbons rather than ten, around 220 grams per mole.
Step five is where vetiver refuses to cooperate, and the refusal is informative. The library names isovalencenol, khusimol and vetiselinenol among hundreds of identified compounds — one of the most complex profiles of any essential oil. None carries this oil the way limonene carries lemon at 60 to 75 per cent. The dominant family answers the question; the dominant compound does not exist.
Steps six and seven follow from size. Sesquiterpenols are among the gentlest groups — no heat, no sunlight caution, standard dilution — and the real caution is aromatic dose rather than skin: one to three drops in a diffuser where most oils call for three to four. Big molecules oxidise slowly, so a vetiver bought today is still vetiver in five years.
Diagram to come
A vertical ladder of five real constituents ordered by molecular weight, lightest at the top, with three columns read across each rung: the molecule and its mass, its family and note position, and how long it survives on a smelling strip. To the right, one continuous arrow from top to bottom carrying the boiling-point range, so mass, note, persistence and boiling point read as a single axis rather than four separate facts. Methyl salicylate sits just above linalool on mass but behaves far more persistently, and the drawing should let that contradiction show rather than hide it.
A molecule leaves a liquid when it is moving fast enough to break away from its neighbours, and how strongly it is held depends on how much surface it presents to them. A fifteen-carbon sesquiterpene presents roughly half again as much surface as a ten-carbon monoterpene, and that alone raises the boiling point by around eighty degrees: limonene boils near 175 °C, beta-caryophyllene near 260 °C. Steam distillation gets around those numbers with a trick — two liquids that do not mix have vapour pressures that add rather than average, so the mixture boils below the boiling point of either one.
The same property runs downstream. Heavy molecules evaporate slowly off a smelling strip, which is what a base note is. They are held longer in a blend, which is what a fixative is. They cross skin more slowly, and they meet oxygen less readily — which is why sesquiterpene-rich oils keep for years while a citrus oil is past its best in one or two. See the endocannabinoid system for the one sesquiterpene with a named receptor.
Wintergreen. Everyone thinks they know this smell — chewing gum, toothpaste, root beer — and files it beside peppermint as a mint. Almost everyone is wrong, and the method catches it at step one.
Gaultheria procumbens, family Ericaceae — the heather family: blueberry, cranberry, rhododendron. It is not a mint, it is not related to a mint, and it shares no chemistry with one. The name has been doing the work of a kitchen-garden herb for a creeping woodland shrub.
Step two: leaf — with a footnote step three makes sense of. The compound this oil is made of does not exist in free form in the living leaf: it is stored bound to a sugar, and production soaks the chopped leaves in warm water for about a day so the plant’s own enzyme releases it before distilling.
Step four is the moment. Methyl salicylate, 98 per cent or more — one compound, effectively the whole bottle. Its family is esters: the one you have been taught to read as sweet, soft and mild. Run the family heuristic here and you get precisely the wrong answer.
Step five explains why, in one structural fact. An ester is an alcohol joined to an acid, and its character follows the acid. Lavender’s esters are built on acetic acid — small, plain, aliphatic. Methyl salicylate is built on salicylic acid, which carries a benzene ring, putting it in a different behavioural class. Nothing about what happens in a body follows from that; it is a fact about structure.
There is a second warning in step five, and it is an absence. At 98 per cent no minor constituents are doing anything. Nothing buffers, nothing softens: whatever the molecule does, the oil does.
Step six, and the library agrees: the heaviest caution wording of any oil in this collection. One drop in a minimum of ten drops of carrier oil. Aromatic and topical only — not internal. Toxic if swallowed, kept out of reach of children. Step seven: 152 grams per mole, boiling near 222 °C — a middle note, persistent on skin long after a monoterpene would be gone.
Four things the name hid and the method found: not a mint, not a mixture, not a gentle ester, not a beginner’s oil.
The large word on the front of a bottle is a commercial name — regional, historical, unregulated, often wrong about the plant. The small italic name underneath is the only field that identifies what was distilled.
A common name tells you what someone decided to call it. A botanical name tells you what it is. If it is not on the bottle or its packaging, the supplier has not told you what you bought.
Three runs, three reasonable outcomes. A method that is right most of the time is only useful if you know when it is not.
Constituent reasoning predicts behaviour, which is narrower than it sounds. It is good at volatility, note position, storage life, blending and the broad shape of a caution. It is poor at aroma in any detail — a smelling strip settles in thirty seconds what no profile settles at all.
A minor constituent can run the oil. Percentage by mass and perceptual weight are different scales: some compounds are detectable at parts per billion while others at ten per cent contribute nothing you can smell. Helichrysum’s italidiones are a trace fraction and the reason helichrysum smells as it does.
Whole-oil behaviour is real and a profile does not capture it. Work comparing a complete oil against its isolated major constituents has repeatedly found the whole behaving differently from the sum of its parts — much of it done in laboratories, on cells in a dish, which is not the same as what happens in a person. A constituent list is a model, not a description. See the oil effect.
And published data is often thin. Many oils have no ranges published anywhere, and those that exist are wide because the reality is.
Which way uncertainty should push you
The common error is to treat an unknown oil as a mild one, because nothing about it has been flagged. Read it the other way round. Not knowing is the stronger reason for caution: dilute further than the reasoning suggests, patch test on the inside of a forearm and leave it a day, use it aromatically before you use it on skin, and keep it away from children, eyes and inner ears.
Product ranges differ by country, so an oil named here may not be on your market’s list — another reason the method matters more than the catalogue.
Keep oils away from eyes and inner ears, and out of reach of children. If you are pregnant, nursing, under medical care or taking medication, speak to your healthcare practitioner before use. Nothing here is intended to diagnose, treat, cure or prevent any disease.
A method read once is a method lost by spring. This one needs ten minutes a week: run the seven steps on an unfamiliar oil, write the prediction down, then check it.
Writing it down first is not a formality: a prediction held in your head rearranges itself to match the answer. Six lines is enough — family, note position, dilution, aroma in three words, two cautions, two partners.
Then open the oils library and mark each line. It is the answer key: botanical name, family, plant part, extraction, constituents, cautions and partners. Expect to be right about note position and dilution long before aroma.
Smell properly while you do it: one drop on a paper strip, read at ten minutes, at an hour, and again next morning — note position, learned by nose rather than by table. Write the oils you use often into your protocols with the family beside each.
Photograph to come
An open notebook on a table, handwritten, a short list of predictions down the left with ticks and crosses beside them, next to three amber bottles turned so their labels are not fully readable and two or three paper smelling strips laid out. Real handwriting, a real kitchen or desk surface, not a styled flat-lay. The point of the picture is that the practice is ordinary and takes ten minutes.
Fifty-two oils a year, ten minutes each. That is how someone ends up able to be handed a bottle and be right about it.
This is the last lesson in the science course. Everything it points at is somewhere the method gets used.
Lesson 12 of 18 · Science