Use three materials for your first accord — this is not a beginner's limit

Ten materials give you 120 possible trios, and you cannot tell any of them apart. Two studies, from 1989 and 1996, found that people identify at most three or four components in a mixture — and that training does not raise the ceiling. Professional perfumers included.

The question that follows the piece on odour thresholds is always the same: "I bought the ten materials. How many go in the first formula?"

Three. The reason is not "start simple": it is that you cannot smell more than that, and neither can a professional.

What the research says

In 1989 Laing and Francis asked 123 subjects to identify the constituents of stimuli made from one to five odorants. Identification was highest for single odours; once the odorants were mixed, almost nobody named the constituents correctly.

There are two readings. Either the capacity to process several simultaneous odours is limited, or odours in a mixture blend into a new odour that keeps few of the characteristics of its parts. For a perfumer both readings mean the same thing.

The decisive study came seven years later. Livermore and Laing wanted to know whether the limit was simply unfamiliarity with the task, so they tested two panels:

a trained sensory panel (10 women, aged 23–43)
the other made up of working professional perfumers and flavourists (8 men, aged 25–55)

Mixtures of up to five odorants were delivered by an air-dilution olfactometer. The result:

Both panels could discriminate and identify only three or four components of a complex mixture, and this capacity could not be increased by training.

The authors concluded the limit is probably imposed physiologically, or by processing constraints.

Hitting that wall is normal. It is the ceiling of human olfaction, and no amount of practice moves it.

What that means for your first formula

The combinatorics:

	Pairs	Trios
Ten materials	45	120
Three materials	3	1

Put all ten in at once and you are holding a bottle you cannot analyse. It may smell good. You will learn nothing reproducible from it, because you cannot attribute anything.

With three you can exhaust the space: A+B, A+C, B+C, A+B+C. Make all four, live with each for three days. When you finish, you understand how those three interact, and that understanding transfers to the next set.

That is the real meaning of "three teaches more than ten": with three, every change you smell can be traced back to a material. Ten leaves you guessing.

How to choose the three

Here the data helps. Our material records include suggested blending partners. Taking the 398,300 of those pairings that resolve to a primary odour family:

	Share
Within the same odour family	29.1%
Across families	70.9%

Seven pairings in ten cross families. Pick three citrus materials and you get orange juice. Reach across families instead and you watch an accord actually form.

The most frequent cross-family pairings:

Pairing	Count
Fruity × floral	10,587
Herbal × floral	9,882
Green × floral	8,919
Floral × citrus	7,565
Floral × balsamic	6,596
Woody × balsamic	6,423

Floral shows up in five of the top six, and there is a reason. Floral materials tend to act as the joint in a formula, connecting the fast citrus and green materials to the slow woods and balsams.

So a practical opening move: one floral in the middle, one lighter than it, one heavier. From the ten starter materials, for example:

Rose as the floral centre
Bergamot above it (citrus, lighter)
Cedarwood underneath (woody, heavier)

That maps onto "floral × citrus" and "woody × floral", two of the high-frequency pairings above.

The procedure

Working from the 10% dilutions in the previous piece:

Equal parts: 1 mL of each. This is the reference; smell it first.
Double each in turn: three bottles, each with one material pushed to 2 mL, so you hear what changes when one voice gets louder.
Halve each in turn: three more bottles, one material down to 0.5 mL apiece.

Seven bottles. Smell each on a strip when fresh, at ten minutes, and the next day. Write it down. I write the time on the tail of each strip as I go; by the next morning a row of unlabelled strips all looks the same.

When those seven are done you will know which material dominates, which only appears within a narrow ratio, and which one you cannot remove without the whole thing collapsing. One complicated bottle teaches none of that.

A common objection

"But commercial fragrances contain dozens or hundreds of ingredients. Nobody can smell those either."

Correct, and that is the point. Most of the materials in a commercial formula are doing quiet structural work: rounding edges, filling gaps in the evaporation curve, keeping cheap synthetics from sounding thin. Nobody is expected to pick them out one by one.

But you have to be able to express something before there is anything to round off. So start with three.

References

Both studies are on PubMed:

D. G. Laing & G. W. Francis, The capacity of humans to identify odors in mixtures, Physiology & Behavior, 46(5), 809–814 (1989). PMID 2628992

A. Livermore & D. G. Laing, Influence of training and experience on the perception of multicomponent odor mixtures, Journal of Experimental Psychology: Human Perception and Performance, 22(2), 267–277 (1996). PMID 8934843

Next up: the evaporation curve. Why "top, heart, base" is not three stages, and how to use substantivity figures to build a curve that does not collapse in the middle.