Fairy Circles

Summary

Millions of bare, regularly spaced circles, 2 to 12 metres across and ringed by taller grass, dot the grasslands at the inland edge of the Namib Desert. Similar circles were found in Western Australia in 2014, and what makes them is still argued over by two camps of scientists.

A good deal is settled. The circles are not permanent: they appear, grow and fill in again over decades. Gas from below, radioactivity, poisoned soil and dead shrubs have been tested and have failed. Two explanations remain: sand termites that kill the grass, or grasses competing so hard for water that they leave regular gaps. In 2022 and 2023 the two camps published papers whose titles state opposite conclusions about the same grasses.

What is documented

The circles. On the deep sands along the eastern, inland edge of the Namib, where rainfall is between 50 and 150 mm a year, the grassland is punctuated by millions of mostly regularly spaced bare circles, 2 to 10 metres or more across, generally with a margin of taller grass (Tschinkel, 2015). They occur in a narrow band from southern Angola to northern South Africa. Cramer and Barger (2013) give diameters of 2 to 12 metres.

The first descriptions. The earliest scientific reference cited by van Rooyen and colleagues (2004) is by Tinley, in 1971 and 1974, who proposed that the circles were fossil termite nests from a wetter period. In 1979 Theron proposed that the remains of dead Euphorbia damarana shrubs poisoned the soil, and marked several dead shrubs with metal pins in the southern Giribes plains of north-west Namibia.

The life cycle. Walter Tschinkel compared satellite images of the NamibRand Nature Reserve taken in October 2004 and December 2008, and found circles appearing and others filling in. From the turnover he estimated lifespans averaging about 41 years: about 24 years for small circles and 43 to 75 years for large ones. A second method, revisiting circles over several years, gave about 60 years (Tschinkel, 2012).

Tested and excluded. Van Rooyen and colleagues (2004) found no raised radioactivity in soil from the circles, and no growth inhibition in soil from beneath E. damarana. In Tschinkel's field experiments, set up in October 2009 and monitored until 2015, an impermeable barrier buried beneath circles to block any gas from below changed nothing; nor did swapping soil between circles and grassland, or adding micronutrients (Tschinkel, 2015). The sources conflict on one point: van Rooyen's bioassays found that soil from circle centres did inhibit plant growth. In March 2020 Stephan Getzin and colleagues relocated Theron's pins after more than 40 years (Getzin and colleagues, 2021). None of the marked spots had become a circle, and grass was growing at the pins.

The termite paper. In March 2013 Norbert Juergens of the University of Hamburg (who also publishes as Jürgens) argued in Science that the sand termite Psammotermes allocerus creates the circles by removing the grass that sprouts after rain. The bare patch then stores water, which sustains the perennial grass at its edge and, through it, the termites. Science ran the paper with an editor's summary headed "Fairies? No, Termites!"

Australia. In 2014 the environmental scientist Bronwyn Bell alerted Getzin to bare circles about 15 kilometres east to south-east of Newman in the Pilbara. They are about 4 metres across and about 10 metres apart, fringed by spinifex grasses, and most had no ant nests or termite mounds (ABC News, 2016). Getzin and ten co-authors reported in 2016 that they share the Namibian pattern but are driven by a different feedback between plants and water. Cramer and Tschinkel (2025) describe the Australian soils as richer in clay and the water flow there as likely to be surface runoff.

The exchange of 2022 to 2024. Getzin, Holch, Yizhaq and Wiegand (2022) excavated grasses that had sprouted inside circles after rain. Within 8 to 9 days they were wilting; by 19 to 20 days they were dead. In the plots that had received rain most recently, the roots of the dead grasses were undamaged in every case, and the authors concluded that water stress, not termites, killed them. Jürgens and Gröngröft replied in 2023 with 14 years of their own measurements: the soil beneath a circle's dry topsoil holds as much moisture as the surrounding grassland or more, the grasses die first at the centre, and termites kill them by damaging their roots. Getzin and Yizhaq answered in 2024 with 400 soil moisture measurements from the 2024 rainy season, showing the topsoil inside circles significantly drier than outside. Seedlings with roots about 10 centimetres long, they argue, cannot reach the moister sand below.

Linyji. In April 2023 Fiona Walsh and colleagues, including Martu co-authors, presented Aboriginal art, narratives and soil excavations suggesting that the bare circles of Australian desert grasslands, called linyji in Manyjilyjarra and mingkirri in Warlpiri, are pavement nests of Drepanotermes harvester termites. Every trench exposed termite chambers. The paper itself states that the exact origins of the circles are unclear. In January 2024 Getzin and colleagues replied that the Newman circles are extremely regular and differ from common termite gaps; Walsh and colleagues answered that termite gaps can be regular too.

The global survey. In September 2023 Emilio Guirado and six co-authors, using a deep-learning model on satellite images, reported "fairy circle-like" patterns at 263 sites in 15 countries on three continents. Cramer and Tschinkel (2025) noted that none had the extreme regularity of the Namibian or Australian circles. In May 2026 Hezi Yizhaq, Getzin and colleagues posted a preprint, not yet peer reviewed: sites they checked on the ground in Madagascar held termite mounds and biocrust, and flagged Namibian sites held Euphorbia vegetation, not fairy circles.

Leading explanations

Sand termites (Juergens and colleagues). Contested. Psammotermes allocerus eats the roots of grass that germinates in the patch, keeping it bare; the bare sand stores water and the ring grass lives on it. Against the hypothesis, Cramer and Tschinkel (2025) note that the termites also occur in the grass between circles, that their occupation of circles fell by 33 per cent over four years of drought while circles persist for decades, and that what constitutes a colony of this termite has not been established. The review calls for experiments that remove termites from circles or add them to grassland; when it was written, neither hypothesis had been tested directly.

Vegetation self-organisation (Getzin, Meron, Cramer, Tschinkel and others). Contested; favoured in the 2025 review. Grasses compete for scarce water: those at a patch's edge draw water from it, the seedlings inside die, and competition spaces the patches regularly. Models built on this reproduce the regular spacing, the effect of aridity on circle size and where new circles form. Cramer and Tschinkel (2025) nonetheless state that the main challenge for the hypothesis is a lack of direct empirical evidence, especially on water movement underground, and that models are not tests. Jürgens and Gröngröft (2023) argue that water cannot move sideways through the sand fast enough. Both camps agree that seedlings inside a circle die within weeks of rain. They disagree about why.

Both together (Tarnita and colleagues, 2017). A modelling proposal, not tested experimentally. Competition between termite colonies sets the large, hexagonal spacing, and plant feedbacks produce smaller-scale patterns between the circles. The authors argued that the multi-scale pattern cannot be explained by either mechanism alone.

Gas seepage, radioactivity, soil toxins, micronutrient shortage and Euphorbia poisoning. Tested and not supported. Cramer and Tschinkel (2025) add that the gas hypothesis lacks a source of gas and cannot account for the regular spacing, and that the distributions of Euphorbia and of fairy circles do not match.

Current status

Partially explained. What the circles are is well described: bare patches with a life cycle of decades, regularly spaced, with water stored beneath the bare sand. Several proposed causes have been tested in the field and excluded. What makes them is not settled. In 2025 Treonis and colleagues, studying soil nematodes along 900 kilometres of the Namib, described the ultimate cause as the subject of an ongoing debate.

What would change this record is an experiment rather than another exchange of papers: removing or adding termites, or manipulating soil water, and following the circles for years. Whether the Newman circles are the same phenomenon as the Namibian ones is a separate open question.

Last reviewed: September 2026. Records are provisional. Where the evidence changes, the entry changes. Found an error? Tell us.