Toxic Woods Chart, Rebuilt: Irritants vs Sensitizers and the Respirator Each One Needs
Toxic Woods Chart, Rebuilt: Irritants vs Sensitizers and the Respirator Each One Needs - A step into the workshop.
Stop Wasting Wood
Generate clear cut diagrams instantly with our new Cutlist Optimizer. Plan your cuts to make your next build easier.
The chart is taped inside the door of half the shops I've visited around Puget Sound. Species on the left, a column of plus signs on the right, cocobolo and yew glowing at the top like a threat. Nobody can tell you where it came from. That bothered me enough to trace it, and the answer turns out to be the point.
The most-copied version isn't undated at all. The text file archived at the University of Rochester states the chart "appeared in American Woodturner June 1990" and was originally posted to rec.woodworking by Bruce Taylor. The Wood Database's chart credits more than a dozen sources, among them Woods and Calnan's "Toxic Woods" in the British Journal of Dermatology, 1976, and the first entry is attributed, literally, to "author unknown." Woodcraft's 28-species watch list, bylined Larry Osborn and updated June 2026, credits no origin at all. The South Puget Sound Woodturners' chart page, from my own corner of the country, holds no data: nine links to other charts, the American Woodturner one among them.
The charts are a game of telephone played over thirty-six years, and the potency ratings outlived whatever they rested on. I've come to believe the useful move is to throw the plus signs out and rebuild the list around two things: what kind of reaction a wood causes, and what that implies for the filter between the dust and your lungs.
Irritant or Sensitizer: The Distinction That Sizes Your Respirator
The Wood Database draws the line cleanly. An irritant produces a reaction shortly after exposure that "tends to go away fairly quickly once you're in the clear." A sensitizer has a latency of hours to months, and "once you develop an allergic reaction, the effects tend to get worse over time." The UK Health and Safety Executive puts the mechanism more bluntly: once sensitised, a worker can suffer severe asthma or dermatitis "after repeated exposure or exposure to lower concentrations."
That last clause is the whole argument. An irritant is a dose problem; cut the dose and you cut the symptom. A sensitizer is a threshold problem, and the threshold moves. The dose that sets you off next year is lower than the one that set you off this year, and it doesn't move back.
The cleanest evidence comes from the Danish furniture industry. Jacobsen, Schaumburg, Sigsgaard and Schlünssen, in Annals of Work Exposures and Health in 2021, tracked roughly 2,000 woodworkers across more than 50 plants over six years. Median inhalable dust fell 40 percent, from 1.0 to 0.6 mg/m³. Daily cough, chronic bronchitis and nasal symptoms all dropped. Cutting the dust by 40 percent fixed the irritation.
The Cancer Risk Lives in Oak and Beech, Not the Exotics Bin
IARC classified wood dust as a Group 1 human carcinogen in its 2012 Volume 100C monograph, for cancer of the nasal cavity, paranasal sinuses and nasopharynx. The dose gradient is steep: the pooled analysis IARC cites gives odds ratios for sinonasal adenocarcinoma of 0.6 at low exposure, 3.1 at moderate, and 45.5 at high. The same monograph finds "strong evidence" for hardwood dust and only a small excess for softwoods. OSHA's Woodworking eTool names oak, mahogany, beech, walnut, birch, elm and ash as hardwoods reported to cause nasal cancer in woodworkers; CCOHS assigns it specifically to beech and oak.
White oak and walnut are what my clients ask for. Oak is the most ordinary board in a Seattle shop, and it's one of the two, with beech, that CCOHS names for nasal cancer. That's the first correction: the long-term hazard is cumulative hardwood dust, and the species you're least afraid of is the one you breathe most.
Rebuilding the List From the Racks You Actually Shop
Here's the list restructured by reaction type, drawn from the Wood Database species pages and CCOHS.
| Wood | Class | What's documented |
|---|---|---|
| White oak (1,350 lbf) | Sensitizer | Eye/skin irritation, asthma-like symptoms; nasal cancer |
| Black walnut (1,010 lbf) | Sensitizer | Eye/skin irritation; on OSHA's nasal-cancer list |
| Western red cedar (350 lbf) | Sensitizer, own TLV | Plicatic acid; asthma, decreased lung function |
| Cocobolo (2,960 lbf) | Sensitizer | "Notoriously allergenic"; asthma-like symptoms, pink-eye, nausea |
| Wenge (1,930 lbf) | Sensitizer | CNS effects, abdominal cramps, splinters that go septic |
| Mansonia (1,390 lbf) | Sensitizer | Nosebleeds, asthma; cardiac poisons in bark and heartwood |
| Teak (1,070 lbf) | Sensitizer | Pink eye, rash, hypersensitivity pneumonitis |
| Purpleheart (2,520 lbf) | Sensitizer | Eye/skin irritation, nausea |
| Ipe (3,490 lbf) | Irritant | Lapachol; headaches, vision disturbance |
| Pacific yew (1,600 lbf) | Irritant, direct toxin | Cardiac effects; nearly all parts of the tree toxic |
| Black locust (1,700 lbf) | Irritant | Eye/skin irritation, nausea less common |
Yew and black locust, fixtures on "particularly toxic" lists, are coded as irritants by the Wood Database and the 1990 chart; nobody calls them sensitizers. And purpleheart, which the Wood Database prices in the "low to medium range for an imported hardwood," is a documented sensitizer, which is exactly why it ends up in beginners' cutting boards.
Then there's western red cedar, our regional softwood. At 350 lbf it's the least threatening board on the bench, and it's the only wood with its own exposure limit: ACGIH sets it at 0.5 mg/m³, half the general value, specifically for its asthma effects per OSHA and CCOHS. Carlsten and colleagues, in PLOS ONE in 2013, followed 40 never-smoking cedar asthmatics 17 years after their last exposure. Forty percent remained in moderate-to-severe impairment, and FEV1 had fallen 1.5 litres since diagnosis. Removal from exposure did not undo it.
On precision: the 2023 review by Baatjies, Chamba and Jeebhay in Current Opinion in Allergy and Clinical Immunology reports 30 percent sensitization for tropical woods against 3 percent for non-tropical. It also states that only one major wood allergen, Trip s 1 from obeche, has ever been formally identified. Every chart that ranks one species +++ and another ++ is making a call with no allergen-level basis. Treat the ranking as folklore and the sensitizer/irritant split as the signal.
The Arithmetic: Protection Factor Times the Limit
OSHA defines a Maximum Use Concentration as the assigned protection factor multiplied by the exposure limit (29 CFR 1910.134). The APF table gives a half mask, including a disposable filtering facepiece, a factor of 10; a full facepiece, 50; a loose-fitting powered air-purifying respirator, 25.
NIOSH recommends 1 mg/m³ total dust and ACGIH's TLV is 1 mg/m³ for all species, per OSHA and CCOHS. OSHA's eTool cites a PEL of 15 mg/m³ total and 5 respirable, but OSHA's own Hazard Recognition page says the 1989 limits were remanded by the Circuit Court of Appeals and are "not currently in force." I design to 1 mg/m³ and treat 15 as a number with no health basis behind it.
Now run that against measured exposure. An HSE-funded study of 22 British woodworking sites, published in Annals of Work Exposures and Health, took 252 measurements from 2014 to 2017 and found a median 8-hour exposure of 1.2 mg/m³ and a 90th percentile of 3.9. Sanding was the worst task at a median 3.2 mg/m³, with two of five sanders over 5.
A half mask at APF 10 against the 1 mg/m³ TLV covers a shop up to 10 mg/m³, comfortably above the sanding median. The same mask against cedar's 0.5 mg/m³ covers only 5 mg/m³, and 6 percent of that study's samples exceeded that. Same mask, same dust, half the headroom, purely because the species is a documented sensitizer. For the exotic sensitizers there is no published species limit at all, so I default to HSE's floor for woodworking: a minimum APF of 20, meaning FFP3, for most machining and for emptying collectors.
What I'd Put on My Face, and Why
The value pick is the GVS Elipse P100 elastomeric half mask, $39.99 at Rockler, with replacement filters at $17.99 a pair. P100 means 99.97 percent minimum efficiency. That's APF 10 for forty dollars, the cheapest legitimate step up from a disposable on the market.
For beards, OSHA prohibits tight-fitting respirators outright on anyone with facial hair crossing the seal, which is why the loose-fitting PAPR at APF 25 exists. The Trend Airshield Pro, $489.99 at Rockler with $74.99 filter pairs, has paperwork that doesn't agree with itself. Rockler's product page and Trend's US site rate it at 98 percent at 0.6 microns; Rockler's own 2019 mask guide says 0.3. Rockler's listing carries the manufacturer's statement that it is "not a RESPIRATOR nor a NIOSH approved device," while Trend's site assigns it "APF10 US Factor." I won't resolve that; in practice it means the device can't satisfy a US respiratory protection program.
The class only works if the fit works. A NIOSH lab study of 25 subjects found N95 facepieces delivering simulated protection factors of 110 to 114 and P100s of 4,571 to 9,420, with nobody below APF 10, under laboratory conditions the authors say may not represent the field. In the field, HSE found only 32 percent of sites had fit tested, and 37 percent of respirator wearers were adequately protected. Buy the mask, then fit it.
At the tool, the only spec that ties an extractor to an exposure limit is European: under EN 60335-2-69, per Festool, Class M passes 0.1 percent of dust and covers "all wood dust," while Class H, at 0.005 percent, is for carcinogenic and pathogenic dust including mould. The US has no equivalent, which is why a $719 Festool CT MIDI I HEPA at Woodcraft lists neither a HEPA percentage nor a dust class.
Spalted, moldy or bark-on stock carries a biological hazard, not a chemical one. Cryptostroma corticale, per the Journal of Occupational Medicine and Toxicology, causes maple bark disease, a hypersensitivity pneumonitis, with spore loads up to 170 million per square centimetre. That's a particle-size problem, so it's P100 or FFP3 with no exceptions, and Class H territory for extraction. If the blank came out of a log with the bark still on, treat it as the most hazardous thing in the shop, whatever species the chart says it is.