Null Results

DTRI publishes null results alongside positive ones. Independent research institutes that hide their nulls cannot be trusted; we make ours visible. The list below catalogs hypotheses that DTRI has tested in the course of the great-circle research program and falsified. The pattern of nulls is itself informative — it narrows the space of viable causal mechanisms and rules out specific overclaims that other researchers have proposed.

This list reflects the current state of the analyses. If new data overturns a null, that change is noted explicitly — see "Nulls overturned by later evidence" below.

Falsified claims and standing nulls

The hypotheses below were tested at DTRI and did not survive falsification. Each entry includes the test statistic and a brief description of the test.

Hypothesis	Result	Description
108° sacred longitudinal spacing	FALSIFIED Z = −1.38	The proposed 108-degree spacing rule for sacred-site placement (Carlson; West) returns a negative Z-score — the observed clustering is below the random-circle expectation, not above it. The rule is falsified.
Astronomical alignments of monuments	NULL	Proposed astronomical-alignment patterns across the corpus do not survive statistical testing.
Monument-orientation Rayleigh test	NULL p = 0.75	Across the orientation distribution of monuments in the corpus, the Rayleigh test for non-uniformity does not reject the null. Orientations are consistent with a uniform distribution.
Ancient trade-network correlation	NULL r = 0.04 p = 0.93	The proposed correlation between great-circle proximity and ancient trade-network density is essentially zero.
South America corridor signal (3 databases)	NULL D = −1.25 / 0.19 / 0.13	Tested on three independent South American archaeological databases — Peru Ministry of Culture (D = −1.25), SAAID (D = 0.19), and Pleiades/p3k14c (D = 0.13). The corridor signal does not replicate in any of the three. The pattern is geographically specific to the Old World.
Anatolia corridor signal (LuwianSiteAtlas)	NULL 0 / 483 sites	Zero of 483 LuwianSiteAtlas sites within 500 km of the proposed corridor. The corridor does not pass through Anatolia.
Desert-edge mechanism	REJECTED mean 134 km	The hypothesis that the corridor pattern is explained by sites clustering along desert edges was tested; mean distance from the proposed desert-edge structure is 134 km, far from the threshold that would support the mechanism.
Desert kites orientation (EAMENA)	NULL p = 0.19 (n = 838)	Across 838 Neolithic desert-kite sites in EAMENA, kite orientations do not show non-uniform structure aligned with the proposed corridor (Rayleigh p = 0.19).
Agricultural-productivity mechanism	NULL p = 0.57	Tested as an alternative explanation for the corridor's site enrichment. Agricultural productivity does not predict on-corridor density above chance.
Lithology mechanism	NULL p = 0.74	Lithological favorability (rock-type substrate) does not predict the corridor pattern. Sites on the corridor are not preferentially on favorable lithology.
Groundwater mechanism	NULL p = 0.37	Water-table depth (Fan et al. 2013 global WTD product) does not predict on-corridor concentration above the Monte Carlo expectation. Tested as a third environmental-favorability null alongside agricultural productivity and lithology.

The three environmental-favorability nulls (agricultural, lithological, groundwater) are jointly informative: they rule out general environmental favorability as the driver of the corridor pattern and strengthen the case for a tectonic–climatic mechanism rather than a generic "habitable land" explanation.

Methodological exclusions

One database was excluded from the analysis after auditing identified a bulk-import bias rather than a real archaeological signal:

• OpenStreetMap — excluded from the great-circle test database after an audit revealed bulk-import inflation in one regional dataset (a satellite-derived tumuli import that did not reflect independent archaeological reporting). The exclusion was a methodological choice to protect the analysis from non-archaeological data, documented in the analysis notes. The canonical monument enrichment of 2.52× (Z = 6.74) is computed on the land-constrained subset of the remaining databases.

Nulls overturned by later evidence

One previously reported null was later overturned by improved database coverage. We list it here rather than the falsified-claims table, so the page accurately represents each null's current status.

Hypothesis	Original / Updated	Description
Arabia corridor signal	Original: 0 / 100 km Updated: Z = 4.70 at 100 km	Initially tested as null using the Pleiades dataset, which has effectively zero coverage of the Arabian Peninsula. With the EAMENA Saudi Arabia subset (14,610 heritage sites), the corridor signal in fact tests positive (Z = 4.70 at 100 km). The original "Arabia null" reflected a database-coverage gap rather than absence of archaeological signal. The corrected analysis is reported in the geological-circumscription manuscript.

What survives

The nulls above narrow the space of viable mechanisms. The findings that remain after this falsification work include:

• Tectonic plate-boundary concentration — the geological-circumscription manuscript reports that the great circle concentrates tectonic plate boundaries at p = 0.011, identifying tectonic–climatic circumscription as the mechanism remaining after the three environmental-favorability nulls.
• Monument enrichment along the corridor — the land-constrained corridor analysis returns 2.52× monument enrichment (Z = 6.74) on N = 10,000 random circles, with 0 / 10,000 random circles matching the observed monument signal.
• Observability gradient — oral-tradition accuracy across 41 knowledge domains correlates with observability at Spearman r = 0.527 (p = 0.0004), with blind-scored subset r = 0.893 (n = 7, p = 0.007).

Each surviving finding is paired with its own data deposit and analysis-code repository. See the papers page for the current portfolio and methods for the audit pipeline.