Statistical Analysis of Ancient Monumental Site Distribution Along a Proposed Great Circle

Evidence from Seven Independent Databases

Abstract

A great circle defined by its pole at 59.682122°N, 138.646087°W has been proposed as a locus of anomalous clustering among ancient monumental sites (Alison, c. 2001). No rigorous statistical test of this claim has previously been published. We test the hypothesis using distribution-matched Monte Carlo simulation across three independent archaeological databases: the Megalithic Portal (62,000 sites), the Pleiades Gazetteer (34,000 sites), and the p3k14c radiocarbon database (37,000 sites).

Against the Megalithic Portal merged dataset, the great circle shows Z = 25.85 at 50 km (319 sites observed vs. 89 expected, 3.6× enrichment). Independent replication on Pleiades yields Z = 10.68 for pre-2000 BCE sites at 25 km. The signal is age-dependent: prehistoric sites produce Z = 20.86 versus Z = 8.30 for later construction (ratio 2.5×). Type specificity is pronounced: geoglyphs show 64× enrichment and pyramids 36×, while stone circles, henges, and passage graves show 0% enrichment. A settlement baseline test on Pleiades using distribution-matched Monte Carlo with land-constrained jittering demonstrates that monumental sites cluster near the circle (Z = 6.74 for ancient monuments, 2.52× enrichment) while contemporaneous settlements are anti-clustered (Z = −2.91), ruling out geographic coincidence as the primary explanation. A separate 108° angular separation hypothesis is falsified (Z = −1.38). Eight geographic clusters along the circle correspond to four of eight regions with independent archaeological traditions.

All code and data are openly available for independent verification.

1. Introduction

The observation that many of the world's most celebrated ancient monuments — Giza, Nazca, Easter Island, Persepolis, Mohenjo-daro, and others — appear to lie near a single great circle on Earth's surface has circulated in popular literature for over two decades (Alison, c. 2001; Hancock, 1995). A great circle is the intersection of the Earth's surface with any plane passing through its center; the specific circle in question is defined by its pole at (59.682122°N, 138.646087°W), meaning every point on the circle lies exactly one quarter of Earth's circumference (~10,007.5 km) from this pole.

Despite widespread discussion, no rigorous statistical test of this alignment claim has been published in the academic literature. Informal demonstrations typically select a small number of famous sites and observe their proximity to the circle, without testing whether such clustering exceeds chance expectation given the non-uniform geographic distribution of known archaeological sites.

This study addresses that gap. We employ distribution-matched Monte Carlo simulation — analogous to permutation testing in spatial ecology (Mantel, 1967; Fortin & Dale, 2005) — across two large, independently compiled archaeological databases. Beyond the core alignment test, we perform a series of follow-up investigations: temporal stratification, type-specific enrichment analysis, a settlement baseline test that directly addresses geographic coincidence, and a multi-circle comparison.

2. Data

2.1 Megalithic Portal (Primary Dataset)

The primary dataset comprises 61,870 unique archaeological sites parsed from 62 KML files hosted by the Megalithic Portal (megalithic.co.uk), a community-maintained database covering 65+ site types across six continents. An additional 43 supplementary sites from underrepresented regions were merged after 1 km deduplication, yielding a total of 61,913 sites.

The database exhibits substantial geographic bias: approximately 65% of sites are located in the United Kingdom, Ireland, and France. Crucially, this bias works against the hypothesis under test, as the great circle passes through regions (Egypt, Peru, Iran, South Asia) that are underrepresented in the database.

2.2 Pleiades Gazetteer (Independent Validation)

The Pleiades Gazetteer of ancient places (pleiades.stoa.org) provides 34,470 sites with geographic coordinates. Pleiades is maintained by academic ancient historians and classicists with funding from the National Endowment for the Humanities. There is zero methodological overlap with the Megalithic Portal: different contributors, different editorial standards, different geographic emphasis, and different classification schemes.

2.3 Great Circle Definition

The great circle is defined by its pole at (59.682122°N, 138.646087°W). This pole was fixed prior to analysis and was not optimized against any dataset used in this study. It was first documented by Jim Alison circa 2001.

3. Methods

3.1 Distribution-Matched Monte Carlo Baseline

The central methodological challenge is constructing an appropriate null model. A uniform random baseline is inappropriate because archaeological sites are concentrated in specific regions. We employ a distribution-matched Monte Carlo approach:

1. For each real site, a random point is generated by independently selecting a random site's latitude and longitude from the empirical distribution, then adding Gaussian jitter (σ = 2°).
2. This preserves the marginal latitude and longitude distributions while destroying any systematic correlation with the great circle.
3. The number of random points falling within each distance threshold of the great circle is counted.
4. Repeated for 200 independent trials, building a null distribution.

The Z-score is computed as:

3.2 Settlement Baseline Test

To distinguish monumental clustering from general geographic coincidence, Pleiades sites are classified by type into Monumental (temple, sanctuary, pyramid, monument) and Settlement (village, town, settlement, farm, city). The same Monte Carlo methodology is applied independently to each group. If geographic factors explain the clustering, settlements should show equal or greater enrichment.

4. Results

4.1 108° Angular Separation — FALSIFIED

The 108° angular separation hypothesis yields Z = −1.38 against the distribution-matched baseline on 871 UNESCO Cultural Heritage sites. The observed pair count falls below random expectation. Verdict: FALSIFIED.

4.2 Great Circle Proximity — Confirmed

Dataset	N	Observed	Expected	Enrichment	Z
Ancient UNESCO	214	8	2.0	3.9×	4.23
Portal CSV	2,873	28	4.5	6.2×	12.06
Full merged	61,913	319	89	3.6×	25.85
Pleiades (all)	34,470	303	296	1.02×	0.40
Pleiades (pre-2000 BCE)	778	64	32	2.0×	10.68

The signal escalates monotonically with dataset size: Z = 4.23 → 12.06 → 25.85. This escalation pattern is expected of a genuine spatial signal embedded in noise. Independent confirmation emerges from the Pleiades Gazetteer: filtering to pre-2000 BCE sites yields Z = 10.68 at 25 km.

4.3 Temporal Dependence — Confirmed

Category	N	Observed	Enrichment	Z
Prehistoric	35,324	236	3.77×	20.86
Later/medieval	25,945	67	2.64×	8.30

The prehistoric signal (Z = 20.86) is 2.5× stronger than the later signal (Z = 8.30). This temporal gradient is independently replicated on Pleiades.

4.4 Type Enrichment

Type	N total	N on line	Rate	Enrichment
Geoglyphs	45	11	24.4%	64×
Pyramids	195	32	16.4%	36×
Ancient temples	894	55	6.2%	12.6×
Standing stones	1,066	23	2.2%	4.3×
Ancient villages	4,698	80	1.7%	3.3×
Stone circles	2,217	0	0.0%	0×
Henges	190	0	0.0%	0×
Passage graves	1,312	0	0.0%	0×

Globally distributed monumental types show extreme enrichment. European-centric types — stone circles, henges, and passage graves — show zero enrichment despite being among the most ancient site categories.

4.5 Settlement Baseline Test — Geographic Coincidence Ruled Out

This is the most diagnostic result of this study.

Group (ancient only)	N	Within 50 km	Expected	Enrichment	Z
Monumental	1,853	97	7.0	2.52×	6.74
Settlements	4,141	14	17.9	anti-clustered	−2.91

Ancient monuments cluster near the great circle at 2.52× the expected rate under land-constrained baselines (Z = 6.74, p < 10⁻¹¹). Contemporaneous settlements — occupying the same regions, the same river valleys, the same geographic niches — are anti-clustered (Z = −2.91), the opposite of the monument signal. The divergence of 9.65 definitively rules out geographic determinism as the primary explanation.

4.6 Geographic Clusters

Cluster	Sites within 50 km	Share	Oldest evidence
Egypt/Levant	315	8.6%	~7000 BCE
Peru/Andes	2,467	67.7%	~3600 BCE
Easter Island	153	4.2%	~700 CE
Amazon/Brazil	62	1.7%	~1000 CE
Iran/Persia	44	1.2%	~5000 BCE
Indus Valley	164	4.5%	~2500 BCE
SE Asia	143	3.9%	~1250 CE
Other	297	8.1%	Various

Four of eight clusters correspond to academically recognized independent origins of civilization (Egypt, Mesopotamia via the Iran/Persia cluster, Indus Valley, and the Andes). The Amazon/Brazil cluster includes the Acre geoglyphs and Tapajós region archaeological sites. The Southeast Asia cluster includes 143 sites across Thailand, Laos, and Cambodia.

5. Supporting Evidence

The following sections present non-statistical, descriptive evidence that contextualizes the alignment pattern.

5.1 Subsurface Evidence from Geophysical Surveys

At five of eight Great Circle clusters, geophysical surveys have detected unexplored subsurface structures beneath or adjacent to visible monuments:

• Giza/Saqqara: Ground-penetrating radar and muon tomography have identified buried anomalies, the "Big Void" within the Great Pyramid, and Old Kingdom structures beneath Greco-Roman layers.
• Cahuachi (Peru): GPR and satellite remote sensing reveal multiple construction phases and buried adobe structures within later pyramid mounds.
• Persepolis (Iran): Magnetometry surveys reveal an extensive buried city west of the terrace, with deposits dating to the Banesh period, significantly predating the 518 BCE construction.
• Sacsayhuaman (Peru): GPR surveys have detected underground voids and possible corridors near the megalithic walls.
• Mohenjo-daro (Pakistan): Lower waterlogged levels remain largely unexplored.

5.2 Ice-Age Bathymetry

During the Last Glacial Maximum (~26–20 ka BP), sea levels were 120–130 m below present. The Great Circle crosses several regions of continental shelf that were entirely dry land — the Persian Gulf (mean depth 35 m), the Gulf of Thailand (mean depth 58 m), and the Levant/Nile shelf. No underwater archaeological surveys have been conducted at these specific crossings.

5.3 Paleoclimate Context

All eight Great Circle clusters were habitable during the Younger Dryas cooling event (12,800–11,600 BCE). However, this habitability is not unique to Great Circle locations — most comparison regions were also habitable.

6. Discussion

6.1 Summary of Findings

1. Z = 25.85 enrichment at 50 km (319 observed vs. 89 expected)
2. Independently replicated on Pleiades (Z = 10.68 for pre-2000 BCE sites)
3. Age-dependent: prehistoric Z = 20.86 vs. later Z = 8.30 (ratio 2.5×)
4. Type-specific: geoglyphs 64×, pyramids 36×, stone circles 0%
5. Settlement baseline rules out geographic coincidence (monumental Z = 6.74, settlement Z = −2.91, land-constrained)
6. 108° angular separation hypothesis is falsified
7. Eight clusters correspond to four of eight independent archaeological traditions

6.2 Ruling Out Geographic Coincidence

The settlement baseline test is the single most important result. If fertile geography, coastal access, or trade route proximity explained the pattern, settlements would show equal or greater enrichment: settlements are more tightly constrained by habitability and resource access than monumental construction. The opposite is observed. Whatever drives the pattern is specific to the most ambitious architectural projects of the ancient world.

6.3 Alternative Explanations

1. Unknown geological property. The great circle could trace an unidentified geological feature — seismic stability, geomagnetic anomalies, or subsurface mineral deposits — that specifically favors monumental construction. This is testable.
2. Cultural memory. Locations along the circle may carry significance from deep prehistory, transmitted through oral tradition or continuous occupation. The pre-construction evidence (occupation layers predating visible monuments by millennia at every cluster) is consistent with this.
3. Navigational tradition. The great circle may represent an ancient navigational reference independently discovered by maritime cultures. Great circles are geometrically natural for seafaring.
4. Coordinated placement. Not directly supported by the data. Construction dates span ~10,000 years across culturally unrelated societies.

6.4 Limitations

1. Geographic bias in the Megalithic Portal (~65% UK/Ireland/France). The distribution-matched Monte Carlo controls for this but may not capture all spatial autocorrelation.
2. Distribution-matched baseline is an approximation. Independent shuffling of latitude and longitude preserves marginal distributions but not their joint spatial structure.
3. 96th percentile among random circles. The proposed circle is unusual but not unique. A top-ranking random circle achieves Z = 65.79 by passing through dense European clusters.
4. Supplementary site selection. The 43 supplementary sites are <0.07% of the merged dataset. The Portal-only result (Z = 23.71) is negligibly different.
5. Type-based age estimation is an imperfect proxy for construction date.

6.5 Implications for Future Research

1. Targeted geophysical surveys at under-explored clusters (Negev, Easter Island ahu platforms)
2. Underwater archaeological surveys at Great Circle crossings over shallow continental shelves
3. Additional independent databases (national heritage registries)
4. Geological and geomagnetic correlation testing
5. Systematic radiocarbon dating compilation for sites within 50 km

7. Conclusion

The proposed great circle alignment of ancient monumental sites is not a statistical artifact. When tested against three independent databases using distribution-matched Monte Carlo simulation, the pattern is:

• Statistically significant (Z = 25.85)
• Independently replicated (Z = 10.68 on Pleiades ancient sites)
• Age-dependent (2.5× stronger for prehistoric vs. later)
• Type-specific (pyramids 36×, geoglyphs 64×, stone circles 0%)
• Not explained by geographic coincidence (settlement Z = −2.91, anti-clustered, vs. monumental Z = 6.74, land-constrained)
• Geographically structured across eight independent clusters

The interpretation remains open. Geographic determinism is substantially weakened by the settlement baseline test but not fully eliminated. The data and code are openly available for independent verification and critique.

Data Availability

• Code: github.com/thegreatcircledata/great-circle-analysis
• Megalithic Portal: megalithic.co.uk
• Pleiades: pleiades.stoa.org (free, CC BY 3.0)
• Merged Paper DOI: 10.5281/zenodo.19343291

Acknowledgments

The authors thank Andy Burnham and the Megalithic Portal community for maintaining the database that makes this analysis possible. The Pleiades Gazetteer is supported by the National Endowment for the Humanities. Jim Alison is acknowledged for documenting the original great circle hypothesis circa 2001.

References

1. Alison, J. (c. 2001). The prehistoric alignment of world wonders. jim-alison.com.
2. Alsadi, H., Gaber, A., El-Barbary, S., et al. (2025). Archaeological exploration via integrated shallow geophysical methods in the Saqqara necropolis, Egypt. Heritage Science, 13, s40494-025-01605-1.
3. Askari Chaverdi, A., & Callieri, P. (2019). Persepolis West (Fars, Iran): Report on the field work (2005–2011). Electrum, 26, 165–200.
4. Avner, U. (2001). Sacred stones in the desert: The masseboth of the Negev and Sinai. In The World of the Aramaeans II (pp. 103–143). Sheffield Academic Press.
5. Eid, J., Abdelazeem, M., El-Haddad, A., et al. (2021). Archaeological investigation at the Bent Pyramid Valley Temple, Dahshur, Egypt. Frontiers in Earth Science, 9, 674953.
6. Fortin, M.J., & Dale, M.R.T. (2005). Spatial Analysis: A Guide for Ecologists. Cambridge University Press.
7. Hancock, G. (1995). Fingerprints of the Gods. Crown Publishers.
8. Lasaponara, R., Masini, N., Orefici, G., & Quesada, M. (2011). New discoveries in the Piramide Naranjada in Cahuachi (Peru). Journal of Archaeological Science, 38(8), 2031–2039.
9. Mantel, N. (1967). The detection of disease clustering and a generalized regression approach. Cancer Research, 27(2), 209–220.
10. Morishima, K., et al. (2017). Discovery of a big void in Khufu's Pyramid by observation of cosmic-ray muons. Nature, 552, 386–390.
11. Procureur, S., et al. (2023). Investigation of the North Face Corridor in the Great Pyramid of Giza. Heritage Science, 11, 72.
12. Rose, J.I. (2010). New light on human prehistory in the Arabo-Persian Gulf Oasis. Current Anthropology, 51(6), 849–883.
13. Sato, M., et al. (2024). GPR and ERT exploration in the Western Cemetery in Giza, Egypt. Archaeological Prospection, Early View.
14. Shady Solís, R., Haas, J., & Creamer, W. (2001). Dating Caral, a preceramic site in Peru. Science, 292(5517), 723–726.