It is tempting to think that diluting odor compounds to the same concentration delivers equally strong stimuli. Yet the concentration in a liquid is different from the concentration carried through the air. Alongside each molecule’s tendency to evaporate, the size of the vial and the way air flows through it also matter.
A study published by Conchou and colleagues in 2026 addressed how to bridge that gap in olfactory experiments. Its main subjects were insect antennae rather than human noses. The aim was to estimate the concentrations actually delivered by a simple odor-delivery device, allowing laboratory responses to be compared with concentrations found in nature.
Airflow changes the balance inside a vial
When an odorant solution is left in a closed vial, molecules moving from the liquid into the air eventually balance those returning to the liquid. A partition coefficient describes how the compound is distributed between the air and liquid at that point. It depends not only on the odorant, but also on the solvent and temperature.
During an experiment, however, air flows through the vial and carries the odor out. Once the initially accumulated, odor-rich air has left, dilution by incoming air competes with replenishment from the liquid. The result can be an initial concentration peak followed by a period at a lower, relatively stable concentration.
The researchers considered three compartments: the liquid, the air inside the vial, and the delivery section leading to the antenna. They calculated the quantities of molecules entering and leaving each. This mass-transfer model estimates concentrations during airflow, which cannot be determined from the closed-vial state alone.
Testing 11 plant volatiles
The study selected 11 compounds with different volatilities, including linalool, α-pinene, and indole. All were dissolved in mineral oil, and their air–oil partition coefficients were measured first.
The delivery apparatus used 1 mL of solution in a 4 mL vial, with air flowing at 200 mL per minute. Odor-laden air passed through tubing and mixed with additional air on its way to the antenna. Concentrations were measured with a calibrated photoionization detector along the apparatus and at its outlet, then compared with the calculations.
Indole and β-caryophyllene did not reach stable concentrations within one minute and were excluded from the analysis fitting steady-state concentrations. For the remaining nine compounds, the model captured the relationship between partition coefficients and concentrations reached during airflow.
In this apparatus, the steady-state concentration of compounds below a certain partition coefficient was about 30% of their closed-vial equilibrium concentration. For highly volatile isoprene, it was about 4%. Even in the same vial, the fraction of the starting concentration maintained differed among compounds.
Using the closed-vial concentration directly overestimated the concentration actually delivered by up to about 30-fold. Discrepancies between the model and measurements were less than tenfold at the source-vial outlet. Including the final dilution with air allowed the model to capture the order of magnitude of concentrations reaching the antenna. It is not a universally precise conversion formula that eliminates even small errors.
The detector’s waveform also needs scrutiny
Another important issue is how quickly an odor starts and stops. When a neural response changes slowly, researchers need to distinguish a property of the neuron from the possibility that the odor itself arrived slowly.
For less volatile compounds, the photoionization detector’s signal rose slowly and tailed off after odor delivery stopped. Measurements using a different type of instrument, a mass spectrometer, showed sharper onsets and offsets for compounds including linalool.
The authors interpret this as the detector itself smoothing the waveform through interactions between the molecules and its internal surfaces. Consequently, they could not use this detector’s waveform to validate the model’s complete concentration time course. Estimating steady concentrations and reproducing moment-by-moment waveforms are different levels of achievement.
Do plant odors interfere with moth pheromone detection?
Using this calibration method, the researchers also examined how (Z)-3-hexenyl acetate, a component of green-leaf odor, affects olfaction in the moth Agrotis ipsilon. Laboratory studies had shown that this compound reduces neural responses to sex pheromone, but the question was whether the concentrations involved were plausible outdoors.
Electrical responses of whole antennae were measured in ten antennae from each sex. Separately, pheromone-sensitive neurons were recorded in 20 males; testing at all concentrations was completed in 16. Stimuli were delivered after concentrations had stabilized, avoiding the initial odor peak.
Clear whole-antenna responses occurred at 160 and 1,600 ppb among the tested concentrations. Here, ppb means parts per billion by gas volume. Neural responses to pheromone were also significantly reduced when these two concentrations of the green-leaf compound formed the background. However, the neurons themselves clearly responded to that plant compound only at 1,600 ppb. Whole-antenna responses and responses of specific neurons are not the same.
Meanwhile, across 16 publications measuring air in natural environments and agricultural fields, green-leaf volatiles were often reported at around 0.1–1 ppb, and the values collected by the authors did not exceed 10 ppb. Some measurements could not precisely distinguish individual compounds, but these concentrations are still far below 160 ppb.
The authors therefore suggest that, for this moth and this plant compound, the interference with pheromone detection observed in the laboratory is unlikely in nature. The possibility of locally concentrated odor filaments remains. The conclusion was not extended to actual mating behavior or to every insect and plant volatile.
Recalibrate whenever the apparatus changes
The model also offers guidance for apparatus design. In the calculations, increasing vial diameter or reducing airflow made concentrations during airflow more likely to approach the closed-vial value. Lower airflow, however, came with a trade-off: it took longer for concentrations to stabilize.
The time for which a constant concentration could be maintained was also predicted to vary greatly with compound volatility. How frequently experimental odor sources should be replaced therefore needs to be considered separately for each compound. This concerns the duration of stable concentrations inside an apparatus, not how long perfume lasts on skin.
Broad trends following changes in vial geometry or airflow can be calculated, but fine-scale turbulence and liquid movement are difficult to estimate accurately. The authors emphasize the need for measurement-based calibration after each change.
Summary
The concentration placed in a vial does not by itself determine the stimulus concentration reaching an olfactory organ. Considering evaporation, replenishment, and dilution by air together provides a way to estimate actual delivered concentrations even with a simple apparatus.
The significance goes beyond making results from different laboratories easier to compare. It enables researchers to reconsider whether a laboratory finding, such as plant odors interfering with moth pheromone detection, is also plausible in nature. The study shows why understanding odor effects requires moving from “what was put in” to “what arrived, and how much.”
References
Conchou, L., Trelea, I.-C., Gévar, J., Monsempès, C., Peron, A.-C., Renou, M., Souchon, I., & Lucas, P. (2026). A method to estimate absolute odorant concentration of olfactory stimuli. PLOS One, 21(1), e0337336. DOI: 10.1371/journal.pone.0337336 / Full original article. Original article: © 2026 Conchou et al., CC BY 4.0.
