Beginner perfumer · 82
156 records carry human toxicity data, and 57 of them involve children
· 11 min read
Toxicity fields mostly talk about rats and mice. But 156 records in this database carry a human figure — labelled oral-man (102), oral-child (57), oral-human (67) and oral-woman (1). Of those, 121 report a TDLo (lowest published toxic dose) and 89 a LDLo (lowest published lethal dose). And the best-supplied entries are not obscure chemicals: tea tree oil at 101 suppliers, methyl salicylate at 88, acetic acid at 82, coumarin at 74, camphor at 66, bitter almond oil at 37. All of these are ingestion figures and have nothing to do with the dose you put on skin — but their existence says one thing plainly: these materials have been through an emergency room.
About 4 minutes.
Toxicity fields go numb after a while. Most of them read like this:
oral-rat LD50 > 5000 mg/kg
Rat, mouse, guinea pig, rabbit. After a few hundred records the numbers become wallpaper.
Until you see "oral-man".
156 records
Sweeping the toxicity_oral column across the database for human labels:
| Subject label | Records |
|---|---|
| oral-man | 102 |
| oral-child | 57 |
| oral-human | 67 |
| oral-woman | 1 |
156 records in total carry at least one human figure (a record can carry several labels).
There are 5,185 records with any oral toxicity data, so human data appears on 3.0% of them.
And the kinds of figure split like this:
- TDLo (lowest published toxic dose): 121 records — the lowest dose at which a toxic effect was observed
- LDLo (lowest published lethal dose): 89 records — the lowest dose at which death was recorded
Eighty-nine.
These are not obscure chemicals
Sorted by supplier count:
| Material | Suppliers | Human figure |
|---|---|---|
| Tea tree oil | 101 | TDLo 500 µL/kg |
| Methyl salicylate | 88 | LDLo 355 mg/kg |
| Acetic acid | 82 | TDLo 1470 µg/kg |
| Coumarin | 74 | TDLo 87 mg/kg |
| (±)-Camphor | 66 | TDLo 51 mg/kg |
| ortho-Guaiacol | 61 | LDLo 43 mg/kg |
| Camphor tree bark oil | 50 | LDLo 50 mg/kg |
| Hydrochloric acid | 48 | LDLo 2857 µg/kg |
| Pine oil | 38 | TDLo 4700 mg/kg |
| Bitter almond oil | 37 | LDLo 107 mg/kg |
| Turpentine oil | 35 | LDLo 2857 µL/kg |
| Mace oil | 34 | TDLo 400 mg/kg |
Tea tree oil, methyl salicylate, coumarin, camphor, bitter almond oil — these live in ordinary houses.
(Coumarin and pine oil have their own articles here, and I did not notice at the time that they were on this list.)
What these numbers can and cannot say
All of them are ingestion figures. TDLo and LDLo record what happened after swallowing; what you do in a formula is skin contact. The doses and absorption routes are different and do not convert.
They are not safety ceilings either. A LDLo is the lowest lethal dose on record — the minimum from case reports, not a line. Below it does not mean safe, and above it does not mean fatal.
Their sample is accidents. This kind of data comes from poisoning cases and occupational exposure reports, not designed experiments. So whether a figure exists depends on whether someone was taken to hospital and written up.
Read the 57 child records that way in particular. They exist because those events happened.
So what is it good for
One: it is a signal that a material has been through an emergency room. No human data does not mean safe, but having human data means the material has caused recordable harm in the real world.
Two: orders of magnitude are comparable. Among LDLo values: ortho-guaiacol 43 mg/kg, camphor tree bark oil 50, bitter almond oil 107, methyl salicylate 355. Things an order of magnitude apart should not get the same level of care.
Three: it is a reminder about storage and labelling. These materials sit on your shelf, and people may sit near your shelf. In Running a perfume activity with children, the most valuable community replies were about exactly this.
Four: it cannot be used to frighten anyone. These materials are legal, common and unproblematic in correct use. Between "toxicity data exists" and "cannot be used" sit dose, route and exposure scenario.
What to do
- When you see
oral-manororal-child, check whether it is TDLo or LDLo. They differ substantially. - Do not convert an ingestion dose into a skin dose. Different route.
- Label it, lock it, and never decant into a food container. That is this list's most direct use.
- No human data does not mean safer; it means no citable case.
- To judge a dose in a formula, read the IFRA standards, not the toxicity field. The two answer different questions.
What this doesn't establish
- This is a field sweep, not a toxicological analysis. I counted how the database records things and consulted none of the primary sources.
- The 156 come from matching four strings — oral-man/child/human/woman. Other phrasings are missed, so this is a lower bound.
- I use the conventional definitions of TDLo and LDLo; sources may adjudicate differently.
- One record can carry several human labels, so the four categories sum to more than 156.
- Much of this data is old (the pine oil figure is from 1981), and case-report quality varies.
- I did not look up the IFRA limit for any of these materials. This article does not address formula doses at all.
- "Human data means an emergency room" is an inference. Some figures may come from occupational exposure or clinical studies instead.