Exploring Earth's Oceans Since 1999 · Founded by Blane Perun
3,072 Articles·80+ Countries·25 Years of Ocean Science·Site Map
◆ Trending
25 years exploring Earth's oceans,
coral reefs, and the creatures within.
3,072Articles
80+Countries
1999Founded

Glass Sponges: Deep-Sea Animals With Skeletons of Silica

Glass Sponges — Living Glass From the Deep Sea

Sponges
November 7, 2026
12 min read
Glass sponge facts
Class
Hexactinellida (phylum Porifera)
Species
623 accepted (2012 census); more being described
Skeleton
Six-rayed (hexactine) silica spicules, often fused
Depth
Mostly 200 to 6,000+ m; shallower in caves and fjords
Tissue
Mostly syncytial: one giant multinucleate “cell”
Reefs
British Columbia: 8,500–9,000 years old, mounds up to ~21 m
Longest spicule
Monorhaphis chuni: ~3 m
Lifespan
Reef sponges 220+ years; some estimates in the thousands

Glass sponges are sponges of the class Hexactinellida, animals that build their skeletons from silica, the main ingredient of glass, in the form of spicules with six rays arranged at right angles. They live mainly in cold, deep water, they are built in a way unlike almost any other animal, and off western Canada they construct the only living sponge reefs on Earth, a type of reef that was thought to have vanished with the dinosaurs.

A 2012 global census of sponges in PLOS ONE by Van Soest and colleagues counted 623 accepted glass sponge species, against more than 7,000 demosponges. Glass sponges are mainly restricted to deeper water, from about 200 m to more than 6,000 m, but they reach scuba depths in a few places, such as submarine caves in the Mediterranean and the fjords and shelf of British Columbia. New species are still being described from almost every deep-sea survey. This page explains how the class is organised, how a glass sponge makes glass in near-freezing water, why its body is effectively one giant cell, how Canada’s sponge reefs work, and what is threatening them.

How do you recognise a glass sponge?

Glass sponges are usually pale, from white to cream or orange, and come in sac, vase, cup, blade and branching-tube shapes. The census authors noted one useful negative: they are never encrusting, unlike many demosponges that spread as thin sheets over rock. Many species are roughly cup-shaped and 10–30 cm tall, with a central cavity that opens at the top, sometimes through a sieve-like plate formed by the skeleton. Some reef-forming species grow much larger, and individual sponges on Canadian reefs can stand more than 1.5 m tall.

The definitive feature is microscopic. Every glass sponge builds spicules based on three axes crossing at right angles, giving six rays, which is where the name Hexactinellida comes from. In some species the spicules stay separate and are held together by tissue, so the skeleton falls apart when the animal dies. In others they fuse into a rigid, three-dimensional lattice that survives death intact, the property that allows glass sponges to build reefs.

glass sponge: Glass Sponges: Deep-Sea Animals With Skeletons of Silica
glass sponge

How are glass sponges classified?

The class is split into two subclasses, distinguished by small spicules called microscleres. The Amphidiscophora carry amphidiscs, rod-shaped spicules with an umbrella-like disc at each end. They contain a single living order with three families, and all have skeletons of unfused spicules. The best known are species of Hyalonema, the glass-rope sponges, which anchor themselves in deep-sea mud with a long tuft of glassy fibres; two new Hyalonema species were described from the Indo-West Pacific as recently as 2025.

The Hexasterophora carry hexasters, star-shaped microscleres. One order, the Lyssacinosida, mostly has unfused spicules; it includes the family Euplectellidae, home of the Venus flower basket, and the rossellid sponges, which form dense mats on Antarctic and North Atlantic shelves. Three other orders have fused, dictyonal main skeletons, and it is from these rigid, lattice-building sponges that the Canadian reef-formers come. A 2023 study of glass sponges from New Zealand by Dohrmann, Reiswig, Kelly and colleagues showed how much remains to be found: expanded sampling of a single region added new insights into the diversity, chemistry and evolutionary relationships of the whole class.

How does a sponge make glass?

The spicules are made of amorphous silicon dioxide, glass without a crystal structure, laid down in layers around a central protein filament. How animals deposit so much silica in cold water is still poorly understood. In demosponges, the other major group of silica-making sponges, the leading explanation involves enzymes called silicateins. When Francis and colleagues sequenced the genome of the reef-building cloud sponge Aphrocallistes vastus, published in Royal Society Open Science in 2023, they found no silicatein genes at all. Instead, the growing tip of the sponge showed high activity of other genes associated with mineralisation, notably one called glassin, along with collagens and actins. They concluded that the two groups may make silica by different, independently evolved mechanisms.

That fits a broader picture of sponge evolution. A dated sponge family tree published in Science Advances in 2026 by Rossi and colleagues concluded that the earliest sponges, in the early Ediacaran period, had no mineral skeleton, and that silica and calcium-carbonate skeletons evolved independently several times. An earlier preprint version of the same work estimated at least four separate origins of silica spicules. If correct, the glass skeleton of a hexactinellid is a solution its lineage arrived at on its own, which may explain why its chemistry differs from that of demosponges.

The A. vastus genome is small, about 80 million base pairs, and unusual in containing nearly 2,500 genes nested inside other genes, more than any other genome the authors compared. Comparing the growing tip, or osculum, with the main body, they found 419 genes expressed at different levels, giving a first list of candidates for how the sponge extends its glass framework.

A body made of one giant cell

Most animals, including other sponges, are built of separate cells, each with its own nucleus. Most of the soft tissue of a glass sponge is instead a syncytium: a single continuous mass of cytoplasm containing many nuclei, which forms both the inner and outer layers of the sponge and connects to the remaining ordinary cells through cytoplasmic bridges. A 2006 study by Leys, Cheung and Boury-Esnault of the cave-dwelling Mediterranean glass sponge Oopsacas minuta traced how this happens. The embryo’s cells stay separate up to the 32-cell stage; then small surface cells fuse with one another and large yolky inner cells fuse to form the multinucleated tissue of the larva and adult.

The syncytium has practical consequences. Because the tissue is continuous, electrical signals can spread through the whole sponge, and glass sponges use them to stop their feeding current, a response that other sponges cannot make so quickly. Glass sponges also cannot contract their bodies the way many demosponges can. In 2023 the first complete glass sponge genome, of O. minuta, was published in BMC Biology by Santini and colleagues. It is among the smallest animal genomes sequenced and surprisingly lacks several genes long considered core to animal body-building, including Wnt and several key transcription factors. The authors linked those losses to the peculiar syncytial body plan. The same study assembled the genome of a new species of archaeon, a Thaumarchaeota, that dominates the sponge’s microbial community.

What do glass sponges eat?

Glass sponges are filter feeders that strain bacteria and other very small particles from the water. The reef-building A. vastus, unlike many sponges, has no microbial symbionts and removes little dissolved organic carbon, so it depends almost entirely on the bacteria it catches. A 2018 study in Scientific Reports by Kahn, Chu and Leys found that glass sponge reefs in the Strait of Georgia graze so heavily that they consume seven times more particulate carbon than could be supplied by particles sinking from the surface alone. Stable isotope analysis suggested that the bacteria come from several sources, including rivers and land, the open ocean and sediment stirred up by tidal currents. Bacteria were still found intact in sponge tissue, and labelled carbon remained, for at least 14 days after a meal.

The ability to switch off the feeding current protects the sponge from clogging. At the Fraser Ridge reef in the Strait of Georgia, a 2018 study in Marine Environmental Research by Grant and colleagues recorded 23 arrests of the feeding current during experiments that raised suspended sediment to between 10 and 80 mg per litre. Single arrests lasted about four minutes. Longer “coughing” arrests averaged about half an hour, switching on and off. Natural sediment levels at the reef averaged 4.4 mg per litre, and the sponges stopped feeding at concentrations below those known to be generated by bottom trawls.

The living glass sponge reefs of Canada

Glass sponges once built reefs on a vast scale. Sponge reefs first appear in the Middle Triassic, and in the late Jurassic a discontinuous reef system about 7,000 km long stretched across the northern Tethys and North Atlantic basins, the largest known biological structure ever to have existed on Earth. Its remains form rock outcrops in Europe today. The reefs declined through the Cretaceous as corals and rudist bivalves rose to dominance, and one hypothesis is that the spread of diatoms, which also build skeletons of silica, drew down the silica that glass sponges needed.

Living sponge reefs were unknown until 1987–1988, when geological surveys found four of them in the Queen Charlotte Basin, in Hecate Strait and Queen Charlotte Sound off northern British Columbia. Three more were reported in the Strait of Georgia in 2005, and further reefs have since been found closer to Vancouver, in Howe Sound, and off Washington and south-east Alaska. The northern reefs lie 70–80 km offshore at 165–240 m depth and together cover more than 700 square kilometres of seabed. Radiocarbon dating of reef cores shows that sponges have been building them for 8,500–9,000 years, since soon after the last ice age.

The reefs are built mainly by three species with fused skeletons: the cloud sponge Aphrocallistes vastus, the chalice sponge Heterochone calyx and Farrea occa. When a sponge dies, its glass framework remains, larvae settle on it, and new sponges grow on the skeletons of earlier generations, much as corals build on dead coral. Fine sediment carried by bottom currents is trapped among the skeletons and buries the lower layers. Over thousands of years this produces mounds, or bioherms, up to about 21 m high, and sheets, or biostromes, 2–10 m thick that can extend for kilometres. Rossellid sponges such as the boot sponge Rhabdocalyptus dawsoni also live on the reefs, but their loose skeletons do not persist after death, so they form mats rather than reefs.

The conditions are very particular, which may explain why these reefs are found nowhere else. They grow in glacier-scoured troughs on gently sloping continental shelf, on rock, gravel and boulders where sedimentation is low, dissolved silica is high at 43–75 micromoles per litre, bottom currents run at 0.15–0.30 m per second, and temperatures stay between about 5.5 and 7.3 °C. Rockfish shelter among the sponges, and pregnant and juvenile rockfish have been seen there, suggesting the reefs serve as nurseries. Crabs, shrimps, prawns and krill also gather around them.

623
Accepted species (2012 census)
8,500–9,000 yrs
Age of BC reefs
700 km²+
Queen Charlotte Basin reefs
21 m
Tallest reef mounds

What threatens glass sponges?

Physical damage is the most immediate threat. A glass skeleton cannot bend, and trawl nets, dredges, anchors and even lines and traps dragged across the bottom snap sponges off. Trawl damage on three of the Queen Charlotte Basin reefs has been seen as parallel tracks 70–100 m apart. Because larvae need the glass skeletons of earlier generations to settle on, a flattened area may not be recolonised quickly, and it has been estimated that broken reef could take up to 200 years to recover. Individual reef sponges grow by up to about 7 cm a year and can live at least 220 years. Parts of the reefs are now closed to bottom-contact fishing, and the Hecate Strait and Queen Charlotte Sound reefs are protected as a marine protected area under Canadian law.

Climate change is a slower but potentially more serious threat. In a 2020 study in Scientific Reports, Stevenson and colleagues exposed A. vastus to warming and ocean acidification, alone and combined. Within a month, 50–60 percent of the sponges exposed to warming had stopped pumping water and 10–25 percent showed tissue withdrawal, which was irreversible. Warming and acidification also reduced the stiffness of the skeleton, and warming weakened it. Environmental data suggested that damaging conditions could occur on the reefs at only 0.5 °C above current temperatures, leading the authors to describe climate change as a serious and immediate threat to the reefs.

Glass sponges in Antarctica: boom and bust

On the Antarctic continental shelf, glass sponges dominate many seabed communities in waters shallower than about 300 m. They are often thought of as slow-growing to the point of stasis, but a 2013 study in Current Biology by Fillinger and colleagues showed otherwise. The collapse of the Larsen A ice shelf on the Antarctic Peninsula in 1995 opened seabed that had lain in darkness under floating ice to plankton production for the first time. A survey in 2007 recorded glass sponges and interpreted them as remnants of the former under-ice fauna. When researchers returned to the same transect four years later, glass sponge biomass had doubled and their numbers had tripled after only two favourable growing seasons.

The authors suggested that Antarctic glass sponges may be locked in arrested growth for decades and then undergo boom-and-bust cycles when food arrives, allowing them to colonise new habitat quickly. What triggers their growth and reproduction remains unknown. More generally, ice-shelf collapses along the Antarctic Peninsula have exposed more than 23,000 square kilometres of seabed to surface production, an opportunity that glass sponges appear well placed to exploit.

How long do glass sponges live?

Very long, in some cases. Growth-ring and isotope studies of the giant anchoring spicules of the deep-sea glass sponge Monorhaphis chuni have produced age estimates in the thousands of years, and Antarctic species have also been suggested as candidates for the oldest living animals, though the methods are debated. Those estimates and their caveats are covered in detail on the page about how long deep-water sponges live. For how glass sponges fit into the wider deep-sea community, see deep-sea sponges; for the giant glass sponges of the far south, Antarctic glass sponges; and for glass sponge reefs through deep time, sponge fossils.

Frequently asked questions

What is a glass sponge?

A sponge of the class Hexactinellida, with a skeleton of six-rayed silica spicules and tissue that is mostly one multinucleated mass.

Are glass sponges made of glass?

Their skeletons are amorphous silicon dioxide, the main ingredient of glass, laid down around protein filaments.

Where do glass sponges live?

Mostly in cold deep water from about 200 m to over 6,000 m, with shallower populations in Mediterranean caves, British Columbia and Antarctica.

Where are the glass sponge reefs?

Off British Columbia, Washington and south-east Alaska; the first were found in Hecate Strait and Queen Charlotte Sound in 1987–88.

How old are glass sponge reefs?

Radiocarbon dating shows the British Columbia reefs have been growing for about 8,500–9,000 years.

How many species of glass sponge are there?

A 2012 census counted 623 accepted species; new ones are still being described.

Can glass sponges respond to stimuli?

Yes. Electrical signals spread through their continuous tissue and stop their feeding current, for example when sediment clouds the water.

Sources & further reading
  1. Van Soest R.W.M. et al. (2012). Global diversity of sponges (Porifera). PLOS ONE. PMC3338747.
  2. Francis W.R. et al. (2023). The genome of the reef-building glass sponge Aphrocallistes vastus provides insights into silica biomineralization. Royal Society Open Science. PMC10282587.
  3. Santini S. et al. (2023). The compact genome of the sponge Oopsacas minuta (Hexactinellida) is lacking key metazoan core genes. BMC Biology. PMC10280926.
  4. Leys S.P., Cheung E., Boury-Esnault N. (2006). Embryogenesis in the glass sponge Oopsacas minuta. Integrative and Comparative Biology.
  5. Rossi M.E. et al. (2026). Independent origins of spicules reconcile paleontological and molecular evidence of sponge evolutionary history. Science Advances (preprint 2024, bioRxiv).
  6. Kahn A.S., Chu J.W.F., Leys S.P. (2018). Trophic ecology of glass sponge reefs in the Strait of Georgia. Scientific Reports. PMC5768768.
  7. Grant N. et al. (2018). Suspended sediment causes feeding current arrests in situ in the glass sponge Aphrocallistes vastus. Marine Environmental Research.
  8. Stevenson A. et al. (2020). Warming and acidification threaten glass sponge Aphrocallistes vastus pumping and reef formation. Scientific Reports. PMC7235243.
  9. Fillinger L. et al. (2013). Rapid glass sponge expansion after climate-induced Antarctic ice shelf collapse. Current Biology.
  10. Dohrmann M. et al. (2023). Expanded sampling of New Zealand glass sponges. PeerJ. PMC10149058.
  11. Sponge reef and Hexactinellid, Wikipedia (reef discovery, extent, conditions and fossil record).
BP
Last updated November 7, 2026
Written by
Blane Perun

Blane Perun has been exploring Earth’s oceans and marine life for over 25 years. Founder of TheSea.Org in 1999, underwater photographer, coral aquaculture pioneer, and explorer of 80+ countries.

About Blane Perun →
Coral Aquaculture · The Pioneer Story
Blane Perun, A coral aquaculture pioneer

Built one of the first home reef tanks designed as a multi-zone replica of natural reef conditions. Author of long-form profiles cited by Wikipedia, Quality Marine, and SeaScope. Independent — no advertisers, no sponsored placements, no editorial board outside the dive log.

Widely credited as the pioneer of named-coral culture in the reef-keeping hobby. The original — “Perun's Purple People Eater”, a Protopalythoa morph from his coral farm — became the first hobbyist coral to carry an aquaculturist's name in trade, sparking the now-standard practice of naming individual coral strains across the industry.

Founded1999 Articles3,072 Countries80+ Cited inWikipedia · Quality Marine · SeaScope
Read Blane's full bio →
The Dispatch · Monthly

Field notes from eighty countries worth of ocean.

One monthly email. New long-form reporting, species profiles, and dive reports — no filler, no tracking pixels, no sponsored placements. Ever.

Sign up and get my one-click underwater color-correction actions for Photoshop — the set I run on my own dive shots. Free.

Get the action now + monthly field notes
No spam · Unsubscribe anytime
Underwater reef photo before and after Blane Perun's Photoshop color-correction action — washed-out blue-green on the left, full natural color restored on the right
Same shot · one click