Material guide: castoreum absolute — the composition was redone in 1993 and 1995, and only 8 of the 24 previously reported compounds were confirmed

About 5 minutes.

Short version

• Name: castoreum absolute, CAS 8023-83-4, EINECS 232-427-5
• Source: the occurrence field reads "beaver"; the TSCA definition says "obtained from castor, castoridae"
• Odour type: leathery — leathery, smoky, animal, phenolic, old wood, burnt, resinous
• Strength: high (smell at 10% or less), rank 3
• Substantivity: the field reads "400 hours at 10% in DPG"
• Appearance: dark brown semi-solid to solid; shelf life 24 months; keep away from heat and light
• Suppliers: 23 — the most of the eleven tariff-0510 records
• No FEMA number, and an empty regulatory field
• All three toxicity fields read "Not determined"

Two fields on the same record use different solvents

There is a detail here that is easy to miss:

• odor_substantivity: "400 hour(s) at 10.00 % in dipropylene glycol"
• odor_descriptions: "at 10.00 % in diethyl phthalate. leathery smoky animal phenolic old wood burnt resinous"

The substantivity was measured in DPG, and the odour description was written in diethyl phthalate. One record, two different baselines.

And the solubility field lists isopropyl myristate and water — no alcohol, no DPG.

Three fields, three solvents. This is common with animal extracts: they do not dissolve well in ethanol, so each party used whatever was to hand, and the fields kept the traces.

(How the numeric column drops the substantivity field's qualifiers is taken apart in That "400 hours" for Ambroxan. This record adds another variant: Ambroxan is "> 400 at 10% in DPG", Javanol "400 at 100%", musk xylol "400 at 20% in DPG", and this one "400 at 10% in DPG". Four records, four measurements, one 400.0.)

Compositional history: two-thirds of what was reported did not come back

Castoreum's chemistry has been studied since the 1940s. Tang, Webster and Müller-Schwarze reanalysed North American beaver (Castor canadensis) castoreum in 1993 and 1995 using GC, GC-MS and NMR, confirming identities against synthesised authentic compounds.

Putting the two papers together:

	Previously reported	Confirmed	Newly found
Phenolic compounds	15	5	10
Neutral compounds	9 (Lederer 1949)	3	13
Total	24	8	23

Three quarters were not confirmed.

The papers' own explanation is that those compounds "are either absent or are not volatile enough to be detected by our methods".

That sentence is worth stopping on. It holds two possibilities at once — the earlier work was wrong, or this method could not see them. The papers do not separate the two, and they do not need to: what they report is "with this method, this is what we could see".

(The same logic came up in How much does a natural vary. A constituent list for a natural material looks like a fact, but it is what one team measured with one method in one year.)

The neutrals paper went a step further: to confirm structures and run bioassays they synthesised eight compounds — 6-methyl-1-heptanol, 4,6-dimethyl-1-heptanol, isopinocamphone, pinocamphone, two linalool oxides and their acetates.

Most of the newly found neutral compounds are oxygen-containing monoterpenes — an entirely different chemical family from the phenolics.

Why the beaver makes it

Both papers open on the same point: North American beavers mark their territories with castoreum, and within their own territories they respond with scent marking to experimental marks made of unfamiliar castoreum or of selected single components.

This is a signalling system, not something made to be smelled by us. "In part, the unique odor of castoreum is due to large amounts of phenolic compounds" appears in both abstracts.

It has a FEMA number, but not on this record

Burdock's 2007 safety assessment is about castoreum extract, FEMA 2261 — prepared by direct hot-alcohol extraction of the castor sacs, "added to food as a flavor ingredient for at least 80 years", and regarded as GRAS by both FEMA and the FDA.

But in this database:

Record	Suppliers	FEMA
Castoreum absolute (this article)	23	none
Castoreum absolute replacer	19	none
Castoreum resinoid	8	none
Castoreum liquid	4	2262
Castoreum tincture	2	none
Castoreum resinoid replacer	1	none
Castoreum extract	0	none

The "castoreum extract" Burdock discusses has zero suppliers in this database, and its record does not carry FEMA 2261.

(When I wrote Only 11 records are genuine animal materials last round, I said castoreum absolute carried a FEMA number too. That was wrong and has been corrected — the FEMA entry belongs to castoreum liquid, number 2262.)

The practical reading: do not assume products sharing a name share a regulatory identity. Of seven records under one CAS, one carries a number.

And how far the "castoreum in your vanilla ice cream" legend can be checked in the odour field is a separate article — across eight castoreum records, the word vanilla never appears.

How to use it

• Evaluate at 10% or less, as the field recommends; its strength rank is 3.
• It is leathery type, not musk type. For animalic character it gives leather, smoke and burnt notes — the opposite direction from civetone's clean musk.
• Note that it does not dissolve in ethanol. The solubility field lists only IPM and water. In practice, make a DPG or DEP dilution first.
• Store away from heat and light; shelf life 24 months.
• It is a dark brown semi-solid, so factor the colour into a finished product.
• It has a replacer with 19 suppliers, and castoreum is one of the few animal materials where the original still leads (23 against 19).

What this doesn't establish

• I have not smelled it. The whole article is fields, database-wide counts and three papers.
• Both Tang papers work on male North American beaver (Castor canadensis). Eurasian beaver (C. fiber) composition may differ, and I did not check where commercial castoreum comes from.
• "8 of 24" adds the two papers' numbers together. Their prior-literature scopes differ (the phenolics paper says "reported elsewhere", the neutrals paper points specifically at Lederer 1949). The addition is mine; the papers do not present it that way.
• "Not confirmed" is not "absent". The papers explicitly allow that the compounds may not be volatile enough to detect.
• I could only read Burdock's abstract, so its annual usage and exposure estimates are not visible to me and no such numbers appear here.
• All three toxicity fields are undetermined. Burdock assesses the extract, not this absolute, and the two do not substitute for each other.
• The three-fields-three-solvents observation is read off the field text. I did not ask suppliers why.
• This record has no physical data at all — no melting point, no vapour pressure, no logP, and the appearance field is marked estimated.