SF-0026Nature Is Terrifying
Kīlauea makes hair out of glass, and it blows into people's gutters
Strands of volcanic glass that can be over a metre long and under a millimetre thick, spun from lava the way sticky dough pulls into threads. The USGS advice for finding some is to leave it alone.
Pele's hair that fell at the USGS Hawaiian Volcano Observatory field station during Kīlauea's episode 46, tangled into a tumbleweed where it landed. The strands are volcanic glass, often thinner than a millimetre. The highlighter cap is the scale.
This macro photograph shows Pele's hair, strands of volcanic glass, that had accumulated into a tumble-weed-like structure where it fell at the USGS Hawaiian Volcano Observatory field station near Kilauea Military Camp in Hawaiʻi Volcanoes National Park. A highlighter cap provides a scale.
- Image
- K. Mulliken
- Taken
- 2026-05-05
- Location
- USGS Hawaiian Volcano Observatory field station near Kīlauea Military Camp, Hawaiʻi Volcanoes National Park, Hawaii
- Archive ref.
- USGS media library — IMG_3571
- Licence
- Public domain as a work of the US federal government, corroborated by the item's Sources/Usage statement (licence record)
- The image itself
- Authentic, unaltered
Pele's hair, Kīlauea episode 46, 5 May 2026. K. Mulliken, US Geological Survey Hawaiian Volcano Observatory. Public domain. Cropped.
Verdict
"The volcano makes actual hair out of glass and it blows into town and piles up like tumbleweed."
The claim is supported by primary or authoritative evidence, and nothing of comparable quality contradicts it.
Every element of the claim is stated directly by the US Geological Survey's Hawaiian Volcano Observatory, which monitors the volcano producing it.
IT IS GLASS, AND THE MECHANISM IS DOMESTIC. HVO's own comparison: "Like the thinning strings that form when you pull your hands away from sticky dough, the hair strands form when blobs of lava are torn apart and the fluid lava still connecting them is stretched thin." The strand freezes because "the outer skin of the lava is cooled quickly by the air and forms volcanic glass."
THE DIMENSIONS ARE AS STATED, and by HVO's account of Kīlauea specifically: "These golden strands can be more than a meter (3 feet) long and less than 1 millimeter (0.04 inches) thick."
IT TRAVELS, AND IT ACCUMULATES. The strands "travel with the wind like ash". During Kīlauea's January 2025 fountaining they reached named communities, and where they landed they tangled: "tubular tumbleweeds of Pele's Hair reaching up to 56 centimeters (22 inches)", and elsewhere "up to 60 cm (about two feet) long". During the 2008–2018 Halemaʻumaʻu lava lake, "accumulations about 1 meter (1 yard) wide were typically found on the windward sides of curbs in the closed Halemaʻumaʻu parking lot".
The most extreme travel distance HVO records is not Kīlauea's. After the November–December 2022 Mauna Loa eruption, observers reported strands "at the Mauna Kea visitor center and all the way to Honokaʻa". This file names the volcano each time because the two are different volcanoes on the same island, and the claim under test is about Kīlauea.
THE HAZARD IS REAL AND SPECIFIC. "Both Pele's Hair and volcanic ash are abrasive and can irritate eyes, skin, and respiratory systems." HVO's plainest sentence is the one that settles the tone: "no one wants glass splinters in their eyes or skin that just keep breaking when you try to pull them out."
Verified rather than Mostly True because no element required qualification. The one thing a reader should adjust is the picture, not the facts: this is not glass raining down as shards. It is fibre, fine enough to drift, which is stranger than the version that circulates.
Verified requires 1 primary or authoritative source. This file cites 3.
In short
Lava fountains at Kīlauea produce Pele's hair: fine golden strands of volcanic glass, formed when blobs of molten rock are torn apart and the lava still joining them stretches thin before the air freezes it. Individual strands can exceed a metre. Wind carries them across communities, where in some areas they tangle into tumbleweeds up to 60 centimetres long and settle in gutters. The USGS says residents concerned about their water catchment systems can disconnect the downpipes feeding their tanks while it is falling. Every part of that is ordinary volcanology, and it reads like an invention.
It is made the way you would guess, if you guessed
The US Geological Survey’s Hawaiian Volcano Observatory explains it with dough:
Like the thinning strings that form when you pull your hands away from sticky dough, the hair strands form when blobs of lava are torn apart and the fluid lava still connecting them is stretched thin.
A lava fountain throws molten rock. Blobs separate in flight, and the lava still joining them draws out into a filament. The air chills the outside of that filament faster than it can fall back together — “the outer skin of the lava is cooled quickly by the air and forms volcanic glass” — and what lands is a thread of glass.
More than a metre long, sometimes. Often less than a millimetre thick.
Then the wind takes it
Pele’s hair travels “with the wind like ash”.
When the tradewinds failed between 15 and 18 January 2025, Kīlauea’s fountaining put deposits into Hawaiʻi Volcanoes National Park, the Volcano Golf Course, Volcano Village, Mauna Loa Estates and Ohia Estates — a list of places, not a hazard zone.
Where it lands it does not lie flat. It tangles, into “tubular tumbleweeds of Pele’s Hair reaching up to 56 centimeters (22 inches)”, and in places up to 60 centimetres. During the decade-long lava lake that ended in 2018, “accumulations about 1 meter (1 yard) wide were typically found on the windward sides of curbs in the closed Halemaʻumaʻu parking lot” — inside the park, on kerbs nobody was parking against.
The longest journey HVO records belongs to the other volcano. After Mauna Loa erupted in November and December 2022, people reported strands at the Mauna Kea visitor centre and “all the way to Honokaʻa”, on the north coast. Same island, same material, different vent — worth separating, because it is the detail most likely to be retold as Kīlauea’s.
The part that sounds invented is the ordinary part
Residents on catchment water who are concerned about it getting into their tanks during future fountaining episodes are told:
Disconnect the downpipe that connects your gutter system to your tank when Pele’s Hair is falling.
Volcanic glass fibre, in the gutters, with a route to the drinking water tank, addressed with a piece of practical plumbing advice. HVO also says the amount that fell in January 2025 was “small enough to be handled by most general filtration systems”.
And if you find some, the observatory would rather you left it:
no one wants glass splinters in their eyes or skin that just keep breaking when you try to pull them out
What to correct in your mental picture
Not the facts — the shape. This is not glass raining down as shards or needles. It is fibre: light enough to drift on wind, fine enough to tangle like wool, and it accumulates the way lint does.
And the quantities are small even where the spread is wide. The recent episode left “less than 1 millimeter” of it in the affected communities. Both things are true, and a file that reported only the tumbleweeds would be overselling a real thing — which is the one way to get this story wrong while quoting it accurately.
The evidence
The mechanism, in the observatory's own domestic comparison: "Like the thinning strings that form when you pull your hands away from sticky dough, the hair strands form when blobs of lava are torn apart and the fluid lava still connecting them is stretched thin."
Cited sources:S1
Why the strand survives as glass rather than falling back as lava: "the outer skin of the lava is cooled quickly by the air and forms volcanic glass."
Cited sources:S1
The dimensions are as claimed, and S2 states them of Kīlauea's own fountaining: "These golden strands can be more than a meter (3 feet) long and less than 1 millimeter (0.04 inches) thick." S1 gives the same figures as a general description of the material.
It travels on the wind "like ash". The longest distance HVO records is from a different volcano: after the November–December 2022 Mauna Loa eruption, strands were seen "at the Mauna Kea visitor center and all the way to Honokaʻa", with roadside strands "up to 50 centimeters (20 inches) long".
Cited sources:S1
It gathers rather than dispersing: "tubular tumbleweeds of Pele's Hair reaching up to 56 centimeters (22 inches)", and in some areas tangles "up to 60 cm (about two feet) long".
Cited sources:S2
It reaches inhabited places by name. When the usual tradewinds failed between 15 and 18 January 2025, Kīlauea's fountaining put Pele's hair in "Hawaiʻi Volcanoes National [Park], Volcano Golf Course, Volcano Village, Mauna Loa Estates, and Ohia Estates".
Cited sources:S2
During the 2008–2018 lava lake, "accumulations about 1 meter (1 yard) wide were typically found on the windward sides of curbs in the closed Halemaʻumaʻu parking lot" — inside the national park rather than in a community, which is why this file cites it for scale and not for reach.
Cited sources:S1
The hazard is stated plainly: "Both Pele's Hair and volcanic ash are abrasive and can irritate eyes, skin, and respiratory systems", with gloves, eye protection and a facemask recommended.
Cited sources:S2
And domestically: residents concerned about Pele's hair "getting in your water catchment system during future lava fountain episodes" are told they can "Disconnect the downpipe that connects your gutter system to your tank when Pele's Hair is falling" — though the amount that had fallen was "small enough to be handled by most general filtration systems".
Cited sources:S2
HVO's advice on finding some is not to collect it: "no one wants glass splinters in their eyes or skin that just keep breaking when you try to pull them out" — so "leave the hairs be if you find them."
Cited sources:S1
Evidence that cuts the other way
Findings that complicate or argue against the verdict above. They are listed because leaving them out would make the file look stronger than it is.
The popular picture is wrong even though the claim is right. This is not glass raining down as shards or needles; it is fibre, fine and light enough to drift on wind and tangle like wool. Anyone imagining falling splinters has the hazard roughly right and the phenomenon wrong.
Quantities are small even when the extent is large. The recent episode deposited "less than 1 millimeter (a fraction of an inch) of Pele's Hair in communities". Wide distribution and thin deposition are both true, and reporting only the first would overstate it.
Cited sources:S2
Nothing here establishes a maximum travel distance, and the one long-range observation is not Kīlauea's. "All the way to Honokaʻa" is what HVO reports of the November–December 2022 Mauna Loa eruption — an observation rather than a limit, from a different vent. No source read here gives a figure in kilometres for how far Kīlauea's strands have been carried.
Cited sources:S1
What we still don’t know
- Every source is the Hawaiian Volcano Observatory. That is the right body to ask — it monitors the volcano and publishes the photographs — but it is one voice, and this file has no corroboration from outside USGS. A reader should weigh it as an authoritative single source rather than as agreement between independent parties.
- How far Kīlauea's Pele's hair actually travels. The longest observation HVO reports — strands at the Mauna Kea visitor center and as far as Honokaʻa — belongs to the November–December 2022 Mauna Loa eruption, not to Kīlauea, and is an observation rather than a limit in either case. For Kīlauea, no source read here states a maximum distance or a figure in kilometres; what is established is a named list of communities within roughly the Volcano area.
- Longer-term health effects are not addressed. The sources describe abrasion and irritation of eyes, skin and airways and recommend protective equipment; nothing read here discusses chronic exposure, and this file makes no claim about it.
- HVO places the fall on the named communities in "episode 4, from January 15–18" of the summit eruption that began on 23 December 2024. The same article reports that the eruption "ended its 5th episode" on 23 January 2025 and that "Five fountaining episodes have taken place so far"; the eruption went on well beyond that, and the photograph here is from episode 46, in May 2026. No attempt was made here to reconstruct the full episode chronology.
- This uncertainty formerly read that two of the three sources had no archive capture at all, including both written articles, and that the file's evidence therefore had no durable copy behind it. That was wrong. The Internet Archive holds captures of all three; the earlier check used an endpoint that answers empty for URLs the archive demonstrably has. The durability gap it described did not exist, and the remaining single-source limitation is the one above: every source here is the same observatory.
Corrections
That HVO gives the January 2025 dates as "the fifth of five fountaining episodes"; that "The USGS asks residents to disconnect the downpipes feeding their water tanks while it is falling"; and, in the hook (and in similar words on the card and in the share line), that the strands are "over a metre long and under a millimetre thick", with no qualifier.
HVO places the 15–18 January fall in episode 4; the fifth episode ended on 23 January. HVO says the amount that fell "usually does not require any action", and it offers the downpipe advice to residents who are "concerned about Pele's Hair getting in your water catchment system", for use "when Pele's Hair is falling". The strands "can be more than a meter (3 feet) long" (the lava-fountains article) and are "often less than 1 millimeter (0.04 inches) thick" (HVO's Volcano Watch article on Pele's hairs as a hazard).
In summarising, the source's conditions and hedges were dropped, which turned an optional precaution into an instruction and a possible size into a fixed one. The episode number was misstated. An independent check of the page against the HVO articles found both on 2026-09-25, before the file was used in a social post.
A page that a social post will link to is now re-read against its primary source by a reviewer who did not write it, before the post goes out.
Sources
Every source below was read in full before it was cited. Archive snapshots are recorded at research time, because a file whose sources have evaporated is indistinguishable from one that never had any.
S1Tier 1Read in full
Volcano Watch — Pele's hairs: a beautiful hazard on the Island of Hawaiʻi
US Geological Survey, Hawaiian Volcano Observatory · 2023-03-23
The formation mechanism and the physical description: the sticky-dough comparison, the cooling of the outer skin into volcanic glass, strand length over a metre and thickness under a millimetre, travel on the wind "like ash", metre-wide accumulations against the curbs of the closed Halemaʻumaʻu parking lot during the 2008–2018 lava lake, and the advice to leave found strands alone. Read carefully for which volcano is which. The 50-centimetre roadside strands and the observations at the Mauna Kea visitor center and Honokaʻa are this article's account of the November–December 2022 Mauna Loa eruption, not of Kīlauea. This file cites them as such.
S2Tier 1Read in full
Volcano Watch — Recent lava fountains highlight Pele's Hair hazards
US Geological Survey, Hawaiian Volcano Observatory · 2025-01-23
The Kīlauea half of this file, and the half the claim is actually about. Strand dimensions stated of Kīlauea's own fountaining, the accumulation sizes (tumbleweeds to 56 cm, tangles to 60 cm), the named communities affected between 15 and 18 January 2025 when tradewinds failed, deposition of less than a millimetre in those communities, the abrasive hazard to eyes, skin and airways with the protective equipment recommended, and the catchment-water advice to disconnect downpipes. Not independent of S1: the same observatory's own newsletter. Two articles, one voice.
S3Tier 1Read in full
May 5, 2026 — Episode 46 Pele's hair (USGS media library item)
K. Mulliken · US Geological Survey, Hawaiian Volcano Observatory · 2026-05-05
This file's photograph and its rights record: the item-level public-domain statement and the credit line "USGS photo by K. Mulliken.", which establishes the us-federal-work basis. Also the only source here with a durable archive capture. Also from HVO, so not independent of S1 or S2 either.
How we checked this
What we went looking for, and what we found
Three checks, and the most useful one worked against the file rather than for it.
First, on whether Verified was too strong. Mostly True would apply if any element needed qualification, and none did: mechanism, dimensions, travel, accumulation and hazard are each stated by HVO in terms at least as strong as the claim. The one adjustment is to the reader's mental image, not to a fact, so it belongs in contradicting evidence rather than in the verdict.
Second, against overstating the scale. It would be easy to build a file out of "tumbleweeds of glass in the streets" and stop there. The same source says deposition in communities was "less than 1 millimeter", which is the qualifier that keeps the file honest, and it is recorded rather than omitted. Wide distribution and thin deposition are both true.
Third, and this is the finding: every source is the same observatory. HVO is unquestionably the right body to ask, and it is still one voice. A search for corroboration outside USGS was not productive within the time this file was given, so the limitation is recorded as the first uncertainty rather than disguised by citing three HVO pages as though they were three parties. SF-0025 had genuine cross-agency corroboration and says so; this one does not and says that.
- We searched for evidence against our own conclusion
- We traced where the claim started
- The claim is stated in its strongest form, not a weak version
- The photograph was verified independently of the story
- We checked whether this was settled after the popular version froze
- Image rights were traced to a stated licence
- We checked what we were leaving out
- The verdict sat before publication, and was re-read cold
- Confidence in the verdict
- high
- Verdict assigned
- Most recent authoritative source
- Researched by
- The Strange File
Part of a collection
The Qualifier Is The StoryThe interesting content is in the clause everyone skims.
The Mechanism Was MundaneLandscapes that look impossible, explained by something ordinary and knowable.
Also related
Case Reopened: we quoted a USGS article word for word and still dropped its condition
For seven weeks our Pele's hair file pinned a water-tank instruction on a January 2025 USGS article. The article's sentence was quoted exactly. The words before it, "if you are concerned", were missing.
Three federal answers to "which mountain is biggest", and Everest only wins one
Highest above sea level is Everest. Tallest base to summit is Mauna Kea, by more than a kilometre. Farthest from the centre of the Earth is Chimborazo. NOAA prints all three.