| A | Historical epochs — 12,000 years of drift at Giza (10.5281/zenodo.19390874) |# The Lion Watches The Lion — Regulus and the Eastern Horizon

Eleven ancient sites, five continents, 334,615 computed sky positions — and a published list of every place the alignment was not there.

This is a cross-site archaeoastronomy pipeline. It takes NASA Horizons solar ephemerides and Astropy stellar transforms, and tests whether the star Regulus rose along the architectural axes of eleven monuments — Giza, Gobekli Tepe, Stonehenge, Newgrange, Chichen Itza, Serpent Mound, Nazca and more — at the epochs when those monuments were built.

20 datasets in, 600,987 raw rows, 334,615 above-horizon rows out. Twenty azimuths tested: 11 positive, 9 negative. The negatives are in the README, the paper, and the output tables, at the same prominence as the positives.

Spoke B of the Regulus–Sphinx 2026 research programme.

---

Method: open survey, dual azimuth

Two design decisions do most of the work here, and both cut against finding a result:

No azimuth pre-filter. Every above-horizon Regulus position is retained, not just those near a target. A pre-filtered survey can only ever confirm what it was pointed at; an open survey lets a site come back empty.

Dual azimuth. Every site is tested twice — once against 90 degrees True East (the universal benchmark) and once against its own documented architectural azimuth, sourced from published literature and cross-checked against Google Earth measurement. A site only counts as a hit on the second, site-specific test.

---

Results

The universal benchmark, and why it is not evidence

8 of 11 sites hit exactly 90.000 degrees due-east Regulus rising, computed independently at different epochs and latitudes.

This is stated up front as a null result. Precession guarantees that every Northern Hemisphere site will see Regulus transit 90 degrees at some epoch. It is celestial mechanics, not design, and it functions here as a calibration check that the pipeline is computing correctly — nothing more.

The actual finding

6 of 11 sites show sub-degree Regulus alignment at non-universal, architecturally specific azimuths:

Site	Target azimuth	Tightest match	Deviation	Year
Serpent Mound	300.0° (head)	299.998°	0.002°	9514 BCE
Chinese Pyramids	1.0° (processional N)	1.000°	0.000°	8029 BCE
Chichen Itza	292.0° (west face)	291.990°	0.010°	976 BCE
Pentre Ifan	260.0° (stone window)	260.010°	0.010°	3974 BCE
Stonehenge	51.0° (Heel Stone)	51.666°	0.666°	2401 BCE
Gobekli Tepe	~93° (site E–W extent)	93.000°	0.000°	9497 BCE

The negatives

Published at equal prominence, because a survey that only reports hits is not a survey:

• Gobekli Tepe enclosure azimuths (159–172° SSE) — Regulus never appears there at all. The Gobekli Tepe alignment is an east-rise, not enclosure-specific.
• Stonehenge midwinter sunset (321°) — confirmed negative.
• Newgrange winter solstice passage (137°) — confirmed negative.
• Chichen Itza north staircase (22°) and north face axis (112°) — both confirmed negative.

Gap bridging

The 7999–3000 BCE gap left open by Spoke A is substantially filled: 207,480 rows from six cross-site sources (Chinese Pyramids, Nabta Playa, Newgrange, Pentre Ifan, Serpent Mound, Waun Mawn).

---

The eleven sites

Site	Latitude	Epochs	Rows
Giza	29.9753 N	9000–8900 BCE	5,252
Gobekli Tepe	37.2233 N	9500–8900 BCE	10,504
Nabta Playa	22.5100 N	7600–7400 BCE	5,101
Chinese Pyramids	34.3817 N	8100–4000 BCE	70,968
Waun Mawn	51.9460 N	4000–3200 BCE	70,172
Pentre Ifan	51.9987 N	4000–3200 BCE	42,711
Newgrange	53.6947 N	9999–3100 BCE	16,357
Serpent Mound	39.0255 N	9999–100 BCE	63,680
Stonehenge	51.1789 N	9500 BCE–2000 CE	19,380
Chichen Itza	20.6830 N	1000–500 BCE	24,445
Nazca	14.6921 S	500–400 BCE	6,045

---

Reproducing it

git clone https://github.com/OwainGlyndwr1400/regulus-eastern-horizon.git
cd regulus-eastern-horizon
pip install pandas numpy matplotlib astropy astroquery

python 1.0/spoke_b_data_audit.py    # verify all 17 external sources are present
python 1.0/spoke_b_generate.py      # generate 4 datasets via NASA Horizons (resumable)
python 1.0/spoke_b_normalize.py     # normalise 20 datasets, apply open-survey filter
python 1.0/spoke_b_filter_audit.py  # validate against pre-existing filtered files
python 2.0/spoke_b_synthesis.py     # per-site summaries, gap analysis, 6 plots

Python 3.11+. spoke_b_generate.py uses a checkpoint/resume system, so a dropped Horizons connection does not cost you the run. All site coordinates and architectural azimuths live in one place, 1.0/site_config.py.

Note on data: the 17 external source datasets are not vendored into this repository. They are referenced by path in site_config.py and documented in external_data_inventory.md. The unified output unified_cross_site.csv contains all 334,615 normalised rows from all 20 sources.

---

Outputs

• unified_cross_site.csv — 334,615 rows, one schema across 11 sites
• per_site_summary.csv — tightest-approach table per site
• site_findings.md — dual-azimuth findings, positive and negative
• gap_analysis.md — temporal coverage assessment
• epoch_overlap_matrix.csv — which sites overlap at which epochs
• visualizations/ — deviation bars, Giza vs Gobekli Tepe, polar azimuth plot, drift series, latitude vs deviation

The paper draft classifies every claim as Established, Interpreted, or Speculative, and keeps the three tiers separate. Eleven references, all verified.

The wider programme

Spoke	Focus
A	Historical epochs — 12,000 years of drift at Giza
B (this repo)	Cross-site — 11 archaeoastronomical sites worldwide
C	Full-year rhythm — the complete 2026 annual Regulus cycle
D	March Equinox — Spring versus Autumn comparison
E	Symbolic and textual — ancient source corroboration
F	Atmospheric refraction — refraction-corrected alignment windows

Hub paper: Astronomical Alignment 2026: Regulus and the Great Sphinx (10.5281/zenodo.19164088).

Citation

Ceisiwr, Erydir, and Lumos Aureon. The Lion Watches The Lion: Regulus and the Eastern Horizon — A Cross-Site Analysis of Stellar Alignments at Ancient Monuments. Zenodo, 2026. 10.5281/zenodo.19423144

Licence

CC BY 4.0.

Authors

Erydir Ceisiwr and Lumos Aureon — ORCID 0009-0004-4577-5253