Reading conventions
Because one of the commonest errors in the field is to cite animal data as if they were human data, every data point in this monograph carries a source label: Human data measured in humans · Animal mouse, frog, fly… · Model simulation or assumed value. Yellow-shaded cells in the tables mark places where sources conflict; red-shaded cells mark places where no human data exist.
This monograph reuses the terminology and citation codes of Volumes 1–3 when referring back to them, and presents only new data. All 259 data points with their DOIs are in Appendix A.
Part I. Molecules on the way to the receptor
Between the perfume bottle and the odorant receptor lie three “filters” that most fragrance literature ignores: airflow determines how many molecules reach the olfactory region; the mucus layer determines which molecules can dissolve and gain access to the receptor; and enzymes in the mucus can turn a molecule into a different substance before it is smelled. This part quantifies each filter.
1.1 Airflow: how much air actually reaches the olfactory region?
The olfactory epithelium lies in the olfactory cleft in the roof of the nasal cavity, off the main path of the airstream. The most-cited figure is that about 10% of inhaled airflow passes through this region during quiet breathing. Tracing it back shows that it comes from an enlarged physical model and a computational fluid dynamics (CFD) simulation of one nasal cavity Model; later reviews pooled it with a few other older models into “under 10%” or “under 15%”.
When simulations are run on individual people’s nasal geometries, the picture becomes much broader Model: depending on the nasal vestibule phenotype, the proportion of air reaching the olfactory cleft ranges from about 1.2% to 7.7% at a flow rate of 15 L/min; in three patients with nasal obstruction, the upper third of the nasal cavity — a region wider than the olfactory cleft, so this is an upper bound — received 11.1 ± 7.4% of unilateral airflow. By age, children are at about 13–15% and older adults at about 3–5% (these two figures and the phenotype range were taken via reviews and have not been checked against the originals). One detail illustrates how sensitive the system is: in one model, removing an obstruction at the nasal valve increased total airflow by 23% but increased airflow to the olfactory region by 334%.
In healthy people, most of the resistance (more than half of the total pressure drop) is concentrated around the head of the inferior turbinate. The geometric dimensions of the nasal vestibule correlate moderately with odour detection thresholds (r about 0.3 for l-carvone and 2-phenylethanol, n = 22). In patients with chronic rhinosinusitis, the simulated odorant uptake at the olfactory region correlates r = −0.60 with the detection threshold, and together with other anatomical measures explains 65% of the variance in threshold (n = 29).
1.2 Sniffing behaviour
A classic study in 21 people gave the optimal sniff parameters for odour identification: a flow rate of about 30 L/min, a volume of about 200 cm³ and a duration of about 0.4 seconds — and a single natural sniff provides as much information as seven or more sniffs Human. Per-nostril flow during active sniffing exceeds 300 mL/second.
Sniffing is not a passive behaviour. People adjust their sniff to the very odour being smelled, with a latency of only 160–260 ms. Odour pleasantness explains up to 62% of the variation in sniff volume and 68% of the variation in sniff duration when a quadratic model is used (n = 20). But this relationship is driven almost entirely by unpleasant odours: the more unpleasant the odour, the shallower and shorter the sniff, whereas for pleasant odours pleasantness barely predicts how people sniff. Sniffing behaviour is a protective response rather than a seeking response.
One simulation result Model: across the usual range of sniff flow rates, turbulent versus laminar flow makes no difference to the amount of odorant reaching the olfactory region, because the main resistance lies in the mucosal phase rather than the air phase. According to this model, for the same sniff volume, sniff duration may matter more than intensity — a hypothesis not yet tested directly in humans.
1.3 The nasal cycle
The two nostrils are not equally open. In a study with continuous 24-hour monitoring, all 33/33 people had at least one switch of the dominant side during the day. The mean cycle length was 2.15 ± 1.84 hours (from about 15 minutes to more than 10 hours), longer during sleep (4.5 hours) than during wakefulness (2.0 hours) Human.
One experiment (n = 20, details available only through a secondary source) suggests that the two nostrils perceive the same mixture slightly differently, because rapidly sorbed compounds are favoured at high flow rates and slowly sorbed compounds at low flow rates. However, another study found that detection thresholds of the two nostrils are equal, because people compensate by sniffing for longer through the low-flow nostril.
1.4 Sorption and physicochemistry: why some compounds are easy to smell and others hard
Foundational experiments in frogs in the 1970s showed that the nasal mucosa behaves like a chromatographic column: compounds highly soluble in mucus are retained at the front, while poorly soluble compounds spread evenly Animal. In humans, mucosal solubility is compound-specific and cannot be fully inferred from water solubility.
This explains a result in section 1.1: the correlation between nose shape and odour threshold is seen only with compounds highly soluble in mucus (l-carvone, 2-phenylethanol) and not with d-limonene. Highly soluble compounds are “skimmed off” by the anterior mucosa en route, so the amount reaching the olfactory cleft depends on airflow; poorly soluble compounds arrive regardless of airflow but have difficulty entering the mucus.
Quantitatively, a linear free-energy relationship model based on physicochemistry explains about 76% of the variance in odour thresholds across 353 compounds, with a standard deviation of about 0.82 log units (corresponding to a factor of about 6.6). The same type of equation explains up to 94–95% of eye irritation and nasal pungency. In the interpretation of the group that collected the data, the portion of odour-threshold variance not explained by physicochemistry is consistent with a specific role for receptors — but part of it is measurement noise, since the self-consistency of the threshold data themselves is only about 0.66 log. Trigeminal irritation, by contrast, is almost entirely a matter of physicochemistry.
| Parameter | Value | Data status |
|---|---|---|
| Mucus / air partition coefficient | 10⁻⁶ to 1 (compound-dependent) | Measured directly in humans for very few compounds; models use parameters pooled from many species |
| Diffusion coefficient in mucus / in air | 10⁻⁶–10⁻⁵ / 10⁻²–10⁻¹ cm²/s | Model values |
| Concentration decay length along the nasal cavity | 5–50 cm | Model values |
| Olfactory mucus thickness | ~5 µm (animals) / ~10 µm (model assumption) | No direct measurement yet in the human olfactory cleft |
1.5 Mucus and odorant-binding proteins
Analysis of olfactory cleft mucus in 30 people provides the best quantitative source currently available Human: the amount of mucus per nasal cavity ranged from 7 to 144 mg (median about 37 mg), and fell sharply with age — about 59 mg in people in their twenties against about 24 mg in people in their sixties, that is, about 60% less. This mucus contains 6–21 mg/mL of protein, copper at a concentration of about 75 µM, and the esterase enzyme CES-1 at about five times the concentration found in respiratory mucus.
Olfactory cilia in mammals are immotile because they lack the motor apparatus. The mechanism that renews the mucus layer around the receptors has therefore not yet been quantified.
Odorant-binding proteins in humans are a topic with weaker evidence than the frequency of their mention suggests. Humans have two genes, OBP2A and OBP2B, that are about 90% identical. OBP2A is present in olfactory mucus and has higher affinity for aldehydes and fatty acids (binding constants not yet read), but it is expressed mainly elsewhere in the body. Genetic evidence for its role comes from a single research group working on two small samples (69 and 52 people). There has been no loss-of-function experiment in humans — and in fruit flies, deleting the most abundant odorant-binding protein entirely did not reduce responses Animal.
1.6 Biotransformation: the nose transforms odorants before they are smelled
This is the finding of Part I with the greatest consequences for perfumers. Human olfactory cleft mucus contains enzymes that transform odorants Human. Two kinds of data must be kept apart: the percentages below were measured in vitro on pooled mucus samples; in the breath of living people, metabolites appeared after only a few seconds, fast enough to exceed the perception threshold — but that study did not report proportions and measured nose and mouth together.
Four points emerge from the human data:
- The extent varies widely between compounds. Aromatic esters are hydrolysed by 8% to 99% (p-cresyl acetate almost completely to p-cresol, but anisyl acetate only about 8%), aldehydes are reduced to alcohols or oxidised to acids, whereas tertiary esters — linalyl acetate, terpinyl acetate, isobornyl acetate — are almost not metabolised at all.
- Perception may be perception of the metabolite. With p-cresyl acetate, part of what people smell is actually p-cresol, via receptor OR9Q2.
- Individual differences are large. Esterase activity ranged 5–45% between people (n = 29), about a ninefold difference.
- It may change with age, since mucus volume falls by about 60% — this is an inference; we have not found a measurement of metabolism by age.
Animal data add the mechanism. In rats, inhibiting cytochrome P450 enzymes in the olfactory mucosa increased the electro-olfactogram response to coumarin by 43%, and inhibiting esterases increased the response to isoamyl acetate by 30% — meaning the enzymes normally weaken the signal. Metabolites appear in the mucosa after only about 100 ms Animal. In humans there are as yet no kinetic constants for these enzymes, and the extent of metabolism within a single sniff of about 0.4 seconds has not been measured.
1.7 Orthonasal and retronasal odour
Odour entering through the nostrils (the orthonasal route) and odour rising from the mouth during eating and drinking (the retronasal route) do not give the same experience. This is also where there is a clear conflict between models and measurements:
- A CFD simulation on one human model predicts peak uptake via the retronasal route to be higher by 45–90% Model.
- Electro-olfactogram measurements (n = 10, significant only for 2-phenylethanol) and psychophysics in humans, however, show the retronasal route to be less sensitive Human.
- The two routes activate different brain regions, most clearly with food odours such as chocolate, whereas with non-food compounds the difference is smaller Human.
1.8 Evaporation from a carrier: how well founded are top, middle and base notes?
Industry convention divides top notes (15–30 minutes), middle notes (3–4 hours) and base notes (over 4 hours). This classification is empirical; no standardised vapour-pressure threshold defines it.
The best quantification available is for evaporation from skin: a first-order one-compartment model explains about 65% of the variance in headspace concentration above the skin after one hour (ranging 32–76% depending on the formulation) Human. With a fixative present, the decay follows a single exponential; without one, it is biphasic.
A monograph compiled from 259 sourced facts collected under a common protocol. Practical-implication sections are the compilers’ inferences. Diagrams are newly drawn from published data. Scientific reference material, not medical advice.