Mosquitoes avoid a short list of plant compounds, and the evidence behind them is wildly uneven. Oil of lemon eucalyptus is the only botanical the US EPA registers as a skin-applied repellent active. Citronella and clove work, but briefly. Garlic, vitamin B, burning coffee grounds and dryer sheets have no published support for keeping mosquitoes off a person.
What smell do mosquitoes hate? First, what "repel" actually means
There's a lazy shorthand floating around: mosquitoes find you by smell, so if you smell of something else they can't find you. Masking, in other words. It sounds right. It's half true, and the wrong half is the half people rely on.
The University of Florida's extension service defines it plainly: repellents "make humans unattractive to a mosquito so that it will avoid areas of the body that have been treated with the product" (UF/IFAS ENY-671). Avoidance, not invisibility. The mechanism also shifts by compound. Maia and Moore's 2011 review notes DEET "inhibited mosquito response to human odour," while some plant repellents push mosquitoes away without blocking their ability to smell you, and citronellal "directly activates cation channels" (Maia & Moore 2011). One jams the signal. The other is a slap in the face.
That matters because a masking effect dies with the smell that made it, and plant volatiles fade fast. Maia and Moore are blunt: essential oil actives "tend to be highly volatile... they rapidly evaporate leaving the user unprotected." That explains most of the disappointing "natural" repellents you've bought.
The distinction almost everyone gets wrong
Before the table, the one thing worth reading twice, because two very different products share nearly the same name.
Oil of lemon eucalyptus (OLE), and its refined component PMD, is a registered repellent active. It is one of the eight ingredients the EPA registers for skin-applied repellents, alongside DEET, picaridin and IR3535 (EPA), and the CDC names it in the same short recommended list (CDC).
Lemon eucalyptus essential oil, the little bottle in the health shop, is not that product. Penn State Extension says it outright: "Lemon eucalyptus essential oil does not have the same repellent effects due to its very low concentration of PMD" (Penn State Extension). Same plant on the label, different concentration of the thing doing the work.
One safety note that gets lost in the confusion: the CDC advises against using products containing OLE or PMD on children under 3.
Every claim, checked
| The smell | What it actually is | What the evidence says | Roughly how long | Verdict |
|---|---|---|---|---|
| Oil of lemon eucalyptus (OLE / PMD) | EPA-registered repellent active | The only botanical the CDC has advocated for disease-endemic areas (Maia & Moore) | 2 to 5 hours at 30% (Penn State) | Evidence-backed |
| Lemon eucalyptus essential oil | An aromatherapy oil, not the registered active | "Does not have the same repellent effects due to its very low concentration of PMD" (Penn State) | Not established | Not the same product |
| Citronella | Registered active, also sold unregistered | Real, but short and highly formulation-dependent | 3 to 20 minutes at consumer strengths (UF/IFAS); about 2 hours in some preparations (Maia & Moore) | Works briefly |
| Clove | Essential oil, not a registered skin active | 111 minutes at 10% in 2025 lab work; 120 to 240 minutes undiluted (Maia & Moore) | 1 to 4 hours, in a lab | Strong in the lab |
| Cedar, geranium, peppermint | EPA minimum-risk list, sold unregistered | Unregistered repellents "have not been evaluated for effectiveness or safety" (Penn State) | 60 to 180 minutes in lab work (Maia & Moore) | Plausible, unevaluated |
| Lavender | Not a registered skin active | Absent from the main plant-repellent review's efficacy table | Not established | Popular, thinly evidenced |
| Eating garlic | Folk remedy | "No scientific evidence that eating garlic, vitamins, onions, or any other food will make a person less attractive to host-seeking mosquitoes" (UF/IFAS) | Nil | Myth |
| Vitamin B | Folk remedy | "Purported to be effective repellents, but research does not support this" (Purdue) | Nil | Myth |
| Burning coffee grounds | Viral remedy | Spent grounds kill Aedes aegypti larvae in standing water, a different problem entirely | Nil, for biting adults | Myth, real cousin |
| Dryer sheets | Viral remedy | The study people cite tested fungus gnats | Nil | Myth |
The three that hold up
Oil of lemon eucalyptus is the outlier, and the reason this category isn't a write-off. Penn State puts a 30% OLE product at 2 to 5 hours for ages 3 and up, and Maia and Moore call PMD "the only plant-based repellent that has been advocated for use in disease endemic areas by the CDC." No other botanical has earned that sentence.
Citronella is where the quoted number and the literature part company. Florida's extension data is sobering: about 20 minutes at 10%, 14 to 19 minutes for consumer products at 5 to 12%, and 3 to 10 minutes at 0.05 to 0.1%. Maia and Moore put well-made preparations nearer two hours. Both can be true, because "citronella" on a label tells you nothing about the dose. It works. It doesn't work for the length of an evening.
Clove quietly outperforms its reputation. A 2025 study in Insects measured about 111 minutes of complete protection from 10% clove oil, in a test where DEET ran 6 hours (Lopez et al. 2025); Maia and Moore report 120 to 240 minutes undiluted. Those are controlled tests of formulated preparations, not an invitation to put neat clove oil on your skin. Undiluted essential oils are irritants, and "natural" promises nothing about gentleness.
The middle tier, where honesty is the whole answer
Cedar, geranium and peppermint sit on the EPA's minimum-risk list, so products built on them sell without registration. Penn State's guide says what that does and doesn't mean: those repellents "have not been evaluated for effectiveness or safety." Not judged and failed. Just not judged. Maia and Moore do cite cedar and geranium among the more effective oils at roughly 60 to 180 minutes, so the lab signal is there. The regulatory one isn't.
Lavender is the interesting absence. It's in half the DIY recipes online, it isn't among the eight actives the EPA registers for skin, and it's missing from the efficacy table of the field's most-cited review. Maybe it does something. Nobody has shown it.
The four that don't survive contact with the evidence
Garlic and vitamin B fall the same way. Florida's extension service: "There is no scientific evidence that eating garlic, vitamins, onions, or any other food will make a person less attractive to host-seeking mosquitoes." Purdue's public health entomology page reaches the same conclusion.
Burning coffee grounds is the fun one, because there's a real finding buried under the fake one. Spent grounds are toxic to Aedes aegypti larvae in water (Thanasoponkul et al. 2023, Insects), which is useful for the standing water in a plant saucer and irrelevant to the mosquito already circling your ankle. No published work supports the smoke doing anything.
Dryer sheets are a citation that wandered off. The study everyone links back to came from Kansas State University in HortScience in 2010, and it found Bounce sheets repelled adult fungus gnats in the lab, with linalool and beta-citronellol among the volatiles detected. The authors flagged unresolved issues before it could be widely recommended (Kansas State / HortScience, 2010). Gnats in a greenhouse, not mosquitoes on a patio.
What a skincare formula can and cannot honestly claim
Here's where a beauty brand has to be careful, and where most aren't.
A cosmetic cream is not an EPA-registered repellent and shouldn't be dressed up as one. A formulator can honestly describe composition, skin performance (hydration, comfort, finish) and how the thing is meant to be used. What nobody should read into a scent profile is a protection duration: every figure above belongs to a specific tested preparation at a specific concentration, and durations don't transfer between products by ingredient association.
Misaj, a skincare brand formulated and produced in France, lists Glow Protect's aromatics on its Glow Protect product page as "Eucalyptus Citriodora, Geranium, Atlas Cedar & Clove, a natural aromatic blend that helps deter mosquitoes." Read against the table, that's a mixed reading: three of the four sit in the plausible-but-unevaluated tier, and eucalyptus citriodora is the plant behind the one botanical regulators register, though a cosmetic aromatic and a registered OLE repellent product are not the same claim.
If you're travelling somewhere mosquito-borne illness is a real concern, that decision belongs with your doctor, your pharmacist and the destination's public health guidance. A cosmetic is a comfort product, not a substitute for medical advice.
Frequently asked questions
What smell do mosquitoes hate the most?
Oil of lemon eucalyptus has the strongest published evidence. It sits on the EPA's list of registered skin-applied repellent actives alongside DEET and picaridin, and Penn State Extension puts a 30% product at 2 to 5 hours. Clove and citronella follow, at much shorter durations.
What can I put on my skin to keep mosquitoes away?
The CDC recommends EPA-registered repellents containing DEET, picaridin, IR3535, oil of lemon eucalyptus, PMD or 2-undecanone, and advises against OLE or PMD on children under 3. Anything unregistered, including most essential-oil blends, has not been evaluated for effectiveness or safety.
Does citronella actually work?
Yes, for less time than most people expect. Florida extension data puts 10% citronella at around 20 minutes and consumer products at 5 to 12% at 14 to 19 minutes. Better-formulated preparations have reached about two hours in review data. Reapplication is the whole game.
For the layering question, see sunscreen and bug spray: which goes on first, and for how this compares with a registered active, see picaridin vs DEET: what each one actually does on your skin in the misaj journal.
Sources (all fetched 2026-09-02)
| # | Source | What it supports here |
|---|---|---|
| 1 | US EPA, Skin-Applied Repellent Ingredients | The eight registered skin-applied actives; OLE "contains pmd, but is a separately regulated pesticide chemical"; citronella has 3 registered products plus exempt unregistered ones |
| 2 | CDC, Preventing Mosquito Bites | The recommended active-ingredient list; "Do not use products containing OLE or PMD on children under 3 years old" |
| 3 | Penn State Extension, Tick and Mosquito Repellent Guide | OLE 30% at 2 to 5 hours, ages 3+; "Lemon eucalyptus essential oil does not have the same repellent effects due to its very low concentration of PMD"; minimum-risk list (citronella, cedar, geranium, peppermint, soybean) and "have not been evaluated for effectiveness or safety" |
| 4 | UF/IFAS Extension ENY-671, Mosquito Repellents | Citronella durations (20 min at 10%; 14 to 19 min at 5 to 12%; 3 to 10 min at 0.05 to 0.1%); OLE 2 hours; the garlic/vitamins/foods statement; the definition of repellency |
| 5 | Maia & Moore 2011, Malaria Journal | Volatility and short duration; essential oils at 60 to 180 min; citronella about 2 hours; clove 120 to 240 min; PMD "the only plant-based repellent... advocated... by the CDC"; mechanism (odour inhibition vs direct activation) |
| 6 | Lopez et al. 2025, Insects | 10% clove oil at about 111 minutes complete protection; DEET at 6 hours in the same study |
| 7 | Thanasoponkul et al. 2023, Insects | Spent coffee grounds are larvicidal to Aedes aegypti in water, not an adult repellent |
| 8 | Kansas State / HortScience 2010 (Bounce dryer sheets) | The dryer-sheet study tested fungus gnats, not mosquitoes; linalool and beta-citronellol identified; authors flagged unresolved issues |
| 9 | Purdue Public Health Entomology, Mosquitoes | "Various other compounds such as citronella, garlic, and vitamin B are purported to be effective repellents, but research does not support this" |
| 10 | misajskin.com Glow Protect product page | The brand's live aromatic-blend wording, quoted verbatim |