Why rosehip oil smells like crayons after three months

Why rosehip oil smells like crayons after three months

You bought it in the spring: a small amber bottle, cold-pressed, more expensive per millilitre than anything else on the shelf. You used it twice, put it in the bathroom cabinet, and forgot about it. Now it is September and you have opened it again, and something is wrong. Not dramatic. Just a smell that was not there before — the word most people reach for is crayons, or putty, or old nuts, and once you have noticed it you cannot un-notice it.

The question you actually want answered is: has it gone off, is it my fault, and can I still put it in a face oil. The answers are probably yes, partly, and no.

Rosehip, chia and sacha inchi arrive with the best stories in the cupboard and the highest prices. What they have in common is a large share of one fatty acid — alpha-linolenic acid, the plant omega-3 — and that acid is the reason for both the story and the smell. So the honest question about an omega-3 oil in a cream is not "what does it do" but "how much of it is still there when the customer opens the jar." This piece, part of our series on plant oils by fatty-acid group, is about that question.

First: what "gone off" actually means on a specification

The smell is not a subjective impression you can be talked out of. It has a name in the standards. The Codex Alimentarius standard for named vegetable oils, CXS 210-1999, opens its quality characteristics with a single unglamorous sentence: "The colour, odour and taste of each product shall be characteristic of the designated product. It shall be free from foreign and rancid odour and taste."

That is a food standard, not a cosmetic one, and it is written for the material rather than for your finished serum. But it tells you that a rancid odour is a recognised failure of an oil against its own specification — not a quirk of your nose, and not something a supplier can describe as "natural variation in cold-pressed material."

What it does not tell you is how far the oxidation has gone. Smell is a late signal. By the time the odour is obvious, the primary oxidation products that a peroxide value measures may already have broken down into the secondary ones you are smelling — which is exactly why peroxide value alone is not a complete answer either, and why the measurements are treated properly in the companion article on oxidative stability of oils.

At the bench: a rancid odour is a reason to withdraw the material from your formulating stock, not a reason to use it up quickly in something cheap. An oxidised oil does not become less oxidised inside an emulsion.

Three double bonds, and what each one costs

Alpha-linolenic acid is C18:3 n-3: an eighteen-carbon chain with three double bonds, at carbons 9, 12 and 15. The last of them sits three carbons from the methyl end of the chain, which is all that "omega-3" means — a position, not a virtue.

Between those double bonds sit two carbons that each have a double bond on both sides. Those are the bis-allylic positions, and they are the weak points: the hydrogen atoms attached to them are the easiest in the whole molecule to pull off, and pulling one off is how lipid autoxidation starts. Count them and you have a ranking: oleic acid, with one double bond, has no bis-allylic carbon at all; linoleic acid, with two, has one; alpha-linolenic acid, with three, has two. The gamma-linolenic acid of borage, covered among the GLA oils, also carries two — the difference between the two C18:3 acids is where the bonds sit, not how many there are, and the sources cited here do not compare their oxidation rates head to head.

The size of that effect has been measured, and the numbers are worth reading in full rather than in the single figure that usually gets quoted. Chapman, Kim and Min, in the Journal of Food Science (2009), added 10,000 ppm of oxidised alpha-tocopherol to a mixture of oleic, linoleic and linolenic acids and stored it in the dark at 55 °C, following how much of each acid disappeared. Their reported oxidation ratios of oleic to linoleic to linolenic were 1 : 2 : 3 after five days, 1 : 12 : 26 after thirty days, and 1 : 8 : 16 after thirty-five days.

Read those three rows together and two things become clear. The third double bond does cost something large — an order of magnitude, not a few per cent. And the gap is not a constant you can multiply a shelf life by: it opens and then partly closes again as the faster-oxidising acids run out of themselves to oxidise. One experiment, one set of conditions, one direction of travel.

A second study speaks to shelf life directly, and it was done in Britain. Kochhar and Henry at Oxford Brookes University, in the International Journal of Food Sciences and Nutrition (2009), measured Rancimat induction periods for eight culinary oils at four temperatures — 90, 100, 110 and 120 °C — and extrapolated to storage conditions. For rice bran and hazelnut oil, their prediction matched the storage time the producers printed on the bottle. For most of the rest it did not, and for walnut oil, the least stable of the eight, the predicted shelf life came out 15 to 20 times lower than the printed best-before. They also found that the relationship between unsaturation and the reciprocal of the induction period was linear for the monounsaturated oils and exponential for the polyunsaturated ones. Every oil in this article is in the second group.

At the bench: treat alpha-linolenic acid as a cost line, not a feature line. Every per cent of it you put in a formula is a per cent that has to be protected by the pack, the antioxidant plan and the date on the label, and the protection has to be designed before the oil is bought.

Rosehip against the rest of the shelf, measured

Until recently you had to argue this from structure. Now there is a direct comparison that includes rosehip, and it is the single most useful number in this article.

Fagoaga and colleagues in Antioxidants (2026) ran four oils through the Rancimat method at 373 K — 100 °C — and reported induction periods:

Rosehip was the least stable of the four, by a long way: under the same accelerated conditions it lasted roughly one sixth as long as olive oil and about one ninth as long as jojoba. That is not a shelf-life figure for your product — a Rancimat induction period is a stress test at a temperature no cupboard reaches, and it is especially sensitive to the secondary products that make the smell. It is a ranking, and the ranking is unambiguous.

One more line from the same study is worth carrying. In their electrochemical antioxidant measurement, jojoba had the lowest total antioxidant capacity of the four at 298 K and still had the longest induction period. Resistance to oxidation is a property of the chains, not of the antioxidants riding along with them. Which is the bad news for rosehip: no antioxidant plan is going to turn it into olive oil.

At the bench: if your formula needs a genuinely light, quick oil that will also sit in a warm bathroom for six months, rosehip is not the answer to that brief and no amount of vitamin E will make it one. Use it where its composition is the point and design the product round the fragility.

How much omega-3 is actually in each oil

The alpha-linolenic content differs a great deal between these three oils and — for rosehip especially — between sources of the same oil.

One study used all three as single dietary fats in rats and so reported their compositions side by side under one method (Prostaglandins, Leukotrienes and Essential Fatty Acids, 2016): rosa mosqueta (Rosa rubiginosa) oil at 33 per cent alpha-linolenic acid, sacha inchi at 49 per cent and chia at 64 per cent, against 10 per cent for canola and 1 per cent for sunflower. In that study the chia value was roughly twice the rosehip value.

Other analyses report each oil in its own context.

Three things follow.

The name on the drum does not tell you the species. The 33 per cent figure is Rosa rubiginosa; the 45–51 per cent figures are Rosa canina. Both are sold as rosehip oil in Britain, often without the species anywhere on the listing. A fourteen-point difference in the most fragile component of the oil is not a detail.

The Codex ranges put the exotics in perspective. Low-erucic rapeseed carries 5.0–14.0 per cent alpha-linolenic acid and soya bean 4.5–11.0 per cent, both from Table 1 of CXS 210-1999. Alpha-linolenic acid is not exotic. It is concentrated in these three, which is a different claim.

Even the supplement trade has a consistency problem. The 2024 NMR study that reported 57–65 mol % for chia was mostly looking at something else: across commercial omega-3 supplements it found contents ranging from 8 to over 50 per cent, and in one fish oil sample no detectable DHA at all — which the authors read as possible adulteration. If that is the state of a regulated food category, a cosmetic raw material bought in 250 ml lots deserves at least as much scepticism.

At the bench: write the species into your raw-material specification, not just the common name, and compare every delivered certificate against the published range for that species. A figure well outside it is a question for the supplier before it is an input to a formula.

Rosehip: what is known, and what is not

Cold-pressed rosehip seed oil is noticeably coloured, and a change in the colour or odour of a batch is a reason to investigate it — not proof of anything on its own. Colour neither establishes nor excludes rancidity.

More importantly: we could not locate a controlled trial of topical rosehip seed oil on human skin in the peer-reviewed literature when preparing this article. The claims commonly made for it rest on its composition, on tradition, and on the wider literature about linoleic and linolenic acids in skin — none of which is evidence about a finished cosmetic product. A formulator should present them accordingly, and a British formulator has a specific reason to, which comes up further down.

At the bench: if the reason rosehip is in your formula is a claim you intend to make on the label, stop and check what you actually hold to support it. If the reason is the composition and the feel, you are on solid ground and the rest of this article applies.

Chia and sacha inchi: same chemistry, different profiles

Chia seed oil carries the highest alpha-linolenic proportion of the three — roughly three-fifths — and is correspondingly the one whose oxidation most needs planning for. The 59.54 per cent figure from the 2024 Food Chemistry study came from an optimised microencapsulation run, freeze-drying the oil in whey protein and modified tapioca starch, which is worth knowing because it tells you what that number was for: retention of ALA through a process, not a typical content of an oil you can buy.

Sacha inchi, in the 2025 study of Amazonian seeds, was close to an even split between alpha-linolenic (46.92 per cent) and linoleic acid (38.09 per cent). For both oils the batch certificate matters more than the botanical name.

At the bench: if you are comparing two suppliers' chia, compare figures that were produced the same way. A percentage from a stability or encapsulation study and a percentage from a straight GC on a delivered lot are not the same measurement wearing the same units.

Storage: where the composition is kept or lost

Everything above points one way. Whatever is claimed for an omega-3 oil, the alpha-linolenic acid has to survive from the press, to the drum, to the jar, to the last application — and the enemies are the usual three: oxygen, light and warmth. What follows is formulating and supplier-screening practice, not a universal specification, and it has to be validated on each product against finished-product and final-pack stability data.

• Buy small and buy fresh. Ask for the pressing date and the batch peroxide value, and ask when the peroxide value was measured. Codex quality factors for edible oils set a maximum of 15 milliequivalents of active oxygen per kilogram for cold-pressed and virgin oils, and 10 for refined oils. Those are food-grade ceilings, not cosmetic limits, and peroxide value measures only primary oxidation — but for an omega-3 oil the sensible practice is to prefer batches well below them.
• Store cold, dark, and with as little air as possible. Refrigerated storage of the raw material; amber or opaque containers; decanting into smaller bottles as the drum empties, so the headspace does not grow as the oil does.
• Add at cool-down, at a low percentage, into a base that carries the antioxidant. In a well-built formula the omega-3 oil is an addition, not the oil phase. The oleic and saturated groups carry the structure; the antioxidant plan is written around the omega-3 oil even though it is the smallest component.
• Choose the pack for the oil, not for the shelf. An airless or opaque pack for an omega-3 product is a formulation decision, not a marketing one. A clear glass dropper bottle puts the most fragile ingredient in the formula under the worst conditions available.
• Run the stability study in that pack, at the temperatures the product will actually meet, and follow peroxide value together with a measure of secondary products and the odour.

At the bench: the single cheapest improvement most small formulators can make to an omega-3 product is to halve the pack size. It halves the time the last dose spends in contact with air, and it costs nothing but a conversation with the packaging supplier.

What a British label makes you commit to

This is the part that catches people out, and it is not chemistry.

In Great Britain, Article 19 of the UK Cosmetics Regulation puts a date on your product, and the rule has two branches:

• The date of minimum durability — the date until which the product, stored properly, continues to fulfil its initial function and remains safe — must be preceded either by the hourglass symbol or by the words "best used before the end of", and must be expressed as month and year, or day, month and year, in that order.
• Indication of that date "shall not be mandatory for cosmetic products with a minimum durability of more than 30 months". For those products, you give instead the period after opening for which the product is safe to use, shown by the open-jar symbol followed by a period in months and/or years.

Read that as a formulator rather than as a label designer and it says something uncomfortable. The thirty-month branch is not a default. Choosing it is a statement that your product is durable for more than two and a half years — and for a facial oil built around an oil whose induction period is 4.45 hours at 100 °C, sitting in a clear dropper bottle in a British bathroom, that is a statement someone has to be prepared to defend.

Who defends it is also specified. The government's guidance on making cosmetic products available in Great Britain requires a Responsible Person with a UK established address, and requires the Product Information File to contain the product safety report and "evidence for the cosmetic product's effects" — and to be kept for ten years after the last batch was made available. The date on the bottle is not a guess that gets forgotten. It sits in a file with your name on it for a decade.

At the bench: decide the branch before you design the pack, not after. If your data support twelve months and a 6M period-after-opening, print that and build a small pack round it. Sub-thirty-month products take a dated "best used before the end of" and no open-jar symbol at all — and there is no shame whatsoever in a short date on a genuinely fresh oil. It is the honest version of the story you are already telling.

When it is more honest to leave the omega-3 out

There is a case for it, and formulators should be willing to make it out loud.

If the product will be sold in a clear bottle; if it will be kept warm and used slowly; if the percentage your costing allows is token; or if you cannot run the stability study in the final pack — then the oil may oxidise during intended use unless finished-product testing shows otherwise, and the label is making a promise the jar cannot keep.

In those cases a linoleic oil gives a light feel with fewer bis-allylic carbons — the linoleic group is covered separately — and an oleic oil gives the base. The omega-3 oils belong in small, dark, airless, quickly used products with a real antioxidant plan, and nowhere else.

At the bench: "2 % rosehip for the story" is the version of this ingredient that fails on all three counts — too little to matter, too fragile to survive, and too prominent on the label to defend. Either build the product around it properly or leave it out and say so.

Reading a certificate of analysis for an omega-3 oil

Supplier-screening practice, in the order the numbers matter:

1. Species, in Latin. Rosa canina and Rosa rubiginosa are both "rosehip". Their published alpha-linolenic ranges do not overlap cleanly.
2. Alpha-linolenic acid percentage. Compare it with the published values for that species. A figure far outside them is a question before it is an input.
3. Linoleic acid. Together with the ALA figure it tells you how much of the oil carries bis-allylic carbons — for sacha inchi and for the omega-3-dominant rosehips, that is most of it.
4. Peroxide value, per batch, with the date of testing. The single number that matters most on a delivery in this group. A peroxide value measured at the press six months ago tells you about the press, not about the drum on your bench.
5. Pressing date and storage conditions in transit. Time and warmth before delivery are oxidation the product never gets back.
6. Colour and odour against the grade. A pale "cold-pressed" rosehip, or a chia with a noticeable odour, is a question before it is an ingredient. Codex expects an oil to be free from rancid odour; hold your supplier to that even though the standard is a food one.

The omega-3 group is the one where formulating is mostly stewardship. The chemistry is set the moment the seed is pressed, and everything you do afterwards either preserves it or spends it. Preserve it — small dark pack, real antioxidant plan, an honest date, a stability study on the finished product — and the composition you bought is the composition the customer gets. Spend it, and the omega-3 survives only on the label.

Sources

• Fagoaga C., Moreno A., Fernández-Julián N., Castellano G. Oxidative stability and kinetics of oxidation of rosehip, sunflower, olive and jojoba oils. Antioxidants 15(5), 646, 2026. PubMed 42193267 — induction periods read from Table 4 of the full text.
• Chapman T. M., Kim H. J., Min D. B. Prooxidant activity of oxidized alpha-tocopherol in vegetable oils. Journal of Food Science 74(7), C536–C542, 2009. PubMed 19895457
• Kochhar S. P., Henry C. J. Oxidative stability and shelf-life evaluation of selected culinary oils. International Journal of Food Sciences and Nutrition 60 (Suppl 7), 289–296, 2009. PubMed 19634067
• Rincón-Cervera M. Á. et al. Vegetable oils rich in alpha linolenic acid increment hepatic n-3 LCPUFA… Prostaglandins, Leukotrienes and Essential Fatty Acids 111, 25–35, 2016. PubMed 26995676
• Hasanpour M. et al. 1H NMR-based metabolomics study of the lipid profile of omega-3 fatty acid supplements and some vegetable oils. Journal of Pharmaceutical and Biomedical Analysis 238, 115848, 2024. PubMed 37948777
• Anand V. et al. Encapsulation efficiency and fatty acid analysis of chia seed oil microencapsulated by freeze-drying… Food Chemistry 430, 136960, 2024. PubMed 37531916
• Alejandro Ruiz F. E. et al. Comprehensive characterization and valorization potential of Amazonian Sacha inchi (Plukenetia volubilis L.) seeds… Frontiers in Nutrition 12, 1597300, 2025. PubMed 40469673
• Nikolova K. et al. Characterization of Bulgarian rosehip oil by GC-MS, UV-VIS spectroscopy, colorimetry, FTIR spectroscopy and 3D excitation-emission fluorescence spectra. Molecules 30(19), 3964, 2025. PubMed 41097383
• Codex Alimentarius. Standard for Named Vegetable Oils, CXS 210-1999 (adopted 1999, amended 2015): Table 1 fatty-acid ranges; Appendix §1 quality characteristics (peroxide value, rancid odour); §2.2 definitions of virgin and cold pressed oils. FAO PDF — consulted 22 September 2026.
• Regulation (EC) No 1223/2009, Article 19 (Labelling), as retained GB law. legislation.gov.uk — consulted 22 September 2026.
• Office for Product Safety and Standards. Making cosmetic products available to consumers in Great Britain. gov.uk — consulted 22 September 2026.