Beginner perfumer · 71
Only 11 records in the database are genuine animal materials, and their replacers have more suppliers
· 14 min read
Customs tariff heading 0510 is the one for ambergris, castoreum, civet and musk. Of the 42,225 materials in this database, 11 carry it. Their supplier counts add up to 61; the corresponding replacers add up to 80. Item by item it is sharper: civet absolute 17 against a replacer at 31, ambergris tincture 8 against "ambergris specialty" at 26, castoreum absolute 23 against a replacer at 19. And real deer musk — musk tonquin and musk tonquin absolute — has zero suppliers on both records. A 2014 Neuron paper also gives you a reason not to treat substitution as a neutral move: macrocyclic ketone, nitro, polycyclic and ester musks activate different glomeruli.
About 4 minutes.
The hard part of counting how many animal materials are left in a database is the definition. The natural_synthetic field is vague, and most records whose odor_type reads animal are synthetic.
But one field is not vague: export_tariff.
Heading 0510 is that chapter
HS heading 0510 reads: ambergris, castoreum, civet and musk; cantharides; bile. It is a classification of animal products, not of chemicals.
Of the 42,225 materials in this database, how many carry it?
Eleven.
| Record | CAS | Suppliers |
|---|---|---|
| Castoreum absolute | 8023-83-4 | 23 |
| Civet absolute | 68916-26-7 | 17 |
| Ambergris tincture | 8038-65-1 | 8 |
| Castoreum liquid | 8023-83-4 | 4 |
| Civet (natural) | 68916-26-7 | 3 |
| Civet tincture | 68916-26-7 | 2 |
| Civet resinoid | 68916-26-7 | 2 |
| Castoreum tincture | — | 2 |
| Musk tonquin absolute | 8001-04-5 | 0 |
| Musk tonquin | 8001-04-5 | 0 |
| Castoreum extract | 8023-83-4 | 0 |
Sixty-one suppliers in total, across three animals: beaver, civet and sperm whale (ambergris).
And the two deer musk records are at zero. Two records sharing one CAS (8001-04-5), with no supplier listing either. The database carries it; nobody is selling it.
The replacer side
For the same four, the replacer records are:
| Replacer | Suppliers | Its original |
|---|---|---|
| Civet absolute replacer | 31 | 17 |
| Ambergris specialty | 26 | 8 |
| Castoreum absolute replacer | 19 | 23 |
| Ambergris tincture replacer | 3 | 8 |
| Castoreum resinoid replacer | 1 | — |
Eighty in total, more than the 61 for the originals.
Item by item, civet and ambergris replacers clearly beat their originals (31 against 17, 26 against 8), while castoreum still has more on the original side (23 against 19).
The civet pair is worth remembering: when I matched 87 original-versus-replacer pairs last round, only three had the replacer ahead, and civet had the highest ratio of them (1.82). See The database has 178 "replacers".
What these 11 records look like
They share a shape:
- Almost no physical data. No melting point, no vapour pressure, no logP, and the appearance field is usually marked estimated. Typical of natural extracts.
- The CAS numbers are the "animal-derived extract" kind, not molecular ones. Civet, civet tincture and civet resinoid all share 68916-26-7 — one CAS holding several distinct products, the same situation measured in Ylang ylang oil.
cosmetic_usesmostly reads "perfuming agents", with no retirement flag. On that field they are usable.- Some carry FEMA GRAS and some do not. Civet absolute is FEMA 2319, ambergris tincture 2049, castoreum liquid 2262 — and musk tonquin holds FEMA 2759 while having zero suppliers. Castoreum absolute, the best-supplied of the group at 23, has no FEMA number at all and an empty
regulatoryfield. (Corrected 2026-08-29: this article originally said castoreum absolute was listed too, which was wrong — the FEMA entry belongs to castoreum liquid.)
There is no regulatory prohibition here. Why they receded is not written in this database's fields — that is a supply-chain and ethics question, and the database only records how many companies list them.
Substituting is not a neutral move
If the replacer smelled the same, this would just be a change of supplier. Receptor-level evidence says it is not that simple.
Shirasu and colleagues did something very clear in Neuron in 2014. They traced where musk odours go in the mouse olfactory bulb:
- Muscone activates a small cluster of glomeruli in the anteromedial bulb, and that cluster is highly specific to macrocyclic ketones.
- Synthetic musks with nitro, polycyclic or ester moieties activate distinct but nearby glomeruli.
- They then identified the mouse receptor MOR215-1 and the human musk receptor OR5AN1.
In other words, "musk" is not one slot in the nose. It is several. Natural animal musks belong to the macrocyclic ketone slot; most synthetic replacements belong to others.
(How the four families of synthetic musk differ in structure, regulation and environmental behaviour is Synthetic musks are not one category. That article is about molecules and the environment; this adds that they separate at the receptor too.)
This does not say the replacers are bad. It says that being a replacer is not only about price and provenance — they travel a different route.
What to do
- To count animal materials, use
export_tariff, notnatural_synthetic. A tariff heading is a customs classification and is not vague. - When you see a "civet" or "ambergris" product, check first whether it is the original or a specialty/replacer. In a catalogue they often differ by one word.
- Deer musk has zero suppliers in this database. Whatever is labelled "musk" on the market is almost certainly not it.
- A replacer's record always carries less data than its original. No CAS, no toxicology, no assay.
- Do not assume a replacer smells the same. The receptor-level evidence says macrocyclic ketones and other musks take different routes.
What this doesn't establish
- This is an audit of a tariff field, not a market or ethics study. I counted how many suppliers this database records, not trade volumes, and I have not touched animal welfare or CITES at all.
- Heading 0510 can miss things. Some animal-derived materials may sit under another heading, or have that field empty. "Only 11" is about this one field.
- Zero suppliers does not mean unobtainable. It means none of the suppliers this database covers lists it.
- The five replacers were pulled by name keywords (replacer / specialty); other naming conventions may not have been caught.
- The Shirasu paper is mouse bulb imaging plus receptor identification. For humans it goes as far as identifying OR5AN1 — it is not a human perception experiment.
- "Different glomeruli" does not mean "smells different". That paper establishes that the neural pathways separate, not whether a replacer resembles its original — which I have not measured at all.