Pith. sign in

REVIEW 4 major objections 7 minor 3 references

The orientation of the Amazonian geoglyphs as a clue for their interpretation

T0 review · 4 major / 7 minor · reviewed 2026-08-02 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read The paper claims that the rectangular earthworks of Acre were deliberately oriented to the Sun's annual cycle, with orientation clusters at 68°, 89.5°, and 115°.

desk verdict First orientation dataset of 300+ Amazonian geoglyphs shows a strong non-random signal, but the solar interpretation is under-supported and needs controls. read the letter →

arxiv 2607.14201 v1 pith:ZZIHM3QX submitted 2026-07-15 physics.hist-ph astro-ph.EP

Giulio Magli This is my paper · ORCID
classification physics.hist-phastro-ph.EP
keywords AmazonianearthworksgeoglyphsAcreBrazilarchaeoastronomysolarorientationsatelliteremotesensingcardinaldirectionsceremonialstructures
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper tries to establish that the rectangular Amazonian earthworks of Acre, Brazil, were deliberately oriented according to the Sun's annual cycle, not aligned at random. Using satellite images of 326 square and rectangular structures, the author finds three clusters of orientation: around 68°, 89.5°, and 115°. He interprets 115° as the southern-hemisphere summer-solstice sunrise, 68° as the end-of-May and mid-July Sun that announces the dry season and the second corn harvest, and 89.5° as cardinal east. The author argues this pattern supports a ceremonial, seasonal-calendar function for the structures and offers a way into the builders' worldview given the lack of written records. He explicitly leaves open that the design could stem from practical reasons.

What carries the argument

The method defines the 'azimuth' of each earthwork as the bearing, between its two adjacent sides, that lies closer to due east, so the data occupy the 45°–135° range and can be compared with the solar rising arc (66°–114° at this latitude). Gaussian-kernel density estimation locates peaks in the 326 measured azimuths, and p-value tests measure how unlikely the clustering is under a uniform null. The solar arc and the cardinal direction are the interpretive yardsticks that turn raw bearings into claims about intentionality.

What would settle it

Measure the azimuths of a control set of streams, field boundaries, and roads in the same Acre landscapes; if non-earthwork features cluster at 68°, 89.5°, and 115° as well, the solar interpretation collapses. A second check: lidar under intact forest—if hidden rectangular earthworks have uniformly distributed orientations, the visible sample is a preservation artifact rather than evidence of design.

Watch

Extended reading notes

Core claim

The paper's central claim is that the orientation of the Acre earthworks is non-random and solar. The azimuth distribution of 326 square and rectangular structures shows three clusters at roughly 68°, 89.5°, and 115°; the outer two have strong statistical support and fall inside the solar rising arc, corresponding to dates with agricultural and solstitial significance, while the central cluster indicates a cardinal (due-east) preference. On this basis the paper concludes that at least the regularly shaped geoglyphs were probably ceremonial structures built by a culture that followed the annual cycle of the Sun.

Load-bearing premise

The load-bearing premise is that the observed clustering of orientations reflects deliberate solar alignment rather than a practical or environmental constraint; the paper tests only whether the distribution is uniform, not terrain slope, drainage, or road layout, and the author acknowledges the design may be due to practical reasons.

Editorial extensions

If this is right

  • If the alignments are real, the regular rectangular geoglyphs become among the earliest documented solar-calendrical monuments in Amazonia, predating European contact.
  • The pattern shifts interpretation away from defensive enclosures and toward ceremonial or seasonal gathering places, consistent with the absence of burials and defensive finds inside the structures.
  • The 68° peak indicates an interest not just in solstices but in specific calendrical dates tied to the dry-season onset and the second corn harvest, implying an agricultural calendar.
  • The cardinal peak at 89.5° suggests a symbolic quadripartition of space, a trait shared with many ancient societies.
  • Satellite-image orientation analysis can provide a first-pass cultural hypothesis for earthworks where excavation is not yet possible.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Editorial inference: The decisive control would be to compare the same azimuths with terrain slope, drainage direction, and the orientation of the connecting roads; the solar explanation would be weakened if non-earthwork features share the same peaks.
  • Editorial inference: The 68° cluster may encode a hydrological calendar—the onset of the dry season and river recession—rather than purely solar observation; the paper's own rainfall data point in this direction.
  • Editorial inference: If lidar surveys reveal earthworks under surviving forest with uniformly distributed orientations, the visible sample could be biased; a uniform hidden sample would undercut the claim of deliberate solar design.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 7 minor

Summary. The paper presents a remote-sensing study of the orientation of 324 square/rectangular Amazonian earthworks (geoglyphs) in Acre, Brazil, using Google Earth Pro. The author defines the azimuth of each structure as the side closest to due east and reports a strongly non-uniform distribution (global p=1.5e-15), with peaks at approximately 68°, 89.5°, and 115°. The 115° peak is identified as the summer-solstice sunrise, the 89.5° peak as a cardinal (east) orientation, and the 68° peak as an early-dry-season/agricultural marker. The author argues that these results indicate deliberate solar/cardinal orientation and a ceremonial, rather than practical, function for the structures.

Significance. If the central claim is established, this would be the first archaeoastronomical study of the Acre geoglyphs and a valuable contribution to the debate on their function. The work makes creative use of Google Earth Pro, provides openly available data (supplementary material and kmz file), and applies explicit azimuth definitions. However, the significance of the paper is currently limited by incomplete statistical reporting, an internally inconsistent peak-significance statement, and the absence of any test of non-astronomical explanations. The conclusion that the non-random orientation reflects solar intent is plausible but not yet demonstrated.

major comments (4)
  1. [§2–3] The statistical methodology is not fully specified. The global p-value of 1.5×10⁻¹⁵ is reported without naming the test statistic, the null model, or the binning procedure. Peak significance values are internally inconsistent: the text first claims 'three statistically significant peaks' and then reports 'borderline evidence (p≈0.5)' for the central 89.5° peak. A p-value of 0.5 is not borderline; it indicates no evidence. This is load-bearing because the identification of three solar/cardinal peaks is the paper's principal result. Provide a precise description of the significance test (including null distribution, peak detection, and multiple-testing correction) and report p-values consistently. If the central peak is not significant, exclude it from the definitive results or explicitly downgrade it to a hypothesis.
  2. [§3–4] The paper does not test non-astronomical explanations for the non-uniform azimuth distribution. The author acknowledges (Section 3) that 'this design may be due to practical reasons,' but the only null hypothesis tested is uniformity. The Introduction emphasizes the network of double-bank straight roads connecting earthworks, but the orientations of these roads are never measured. If the rectangular structures are systematically aligned with approaching roads that follow ridges or drainage, the same azimuth peaks could emerge without solar intent. The author should measure road orientations and/or compare the earthwork azimuth distribution with a control sample (e.g., local terrain aspect, modern field boundaries) or with a null model that incorporates such alignments. Without this, the inference from non-uniformity to deliberate solar alignment is not warranted.
  3. [§4] The interpretation of the 68° peak as an early-dry-season/agricultural marker relies on modern climatic and agricultural data (rainfall in Rio Bravo, current double-cropping of corn) without evidence that these conditions existed in the pre-Columbian period. The text itself says 'at least today,' making this a post hoc historical analogy. The peak is only ~2° from the winter-solstice sunrise azimuth, and the two calendar dates (end of May / mid-July) are not independently supported by any archaeological trace. Frame this as a hypothesis rather than a result, and test it against alternative explanations (e.g., horizon-height variations at the sites, or a solstitial interpretation). This is load-bearing for the claim of agricultural interest in the Sun.
  4. [§2] The selection criteria for including parallelograms with internal angles deviating no more than 2° from 90°, and the 'average' used for slightly bended parallelograms, are arbitrary and untested. Since these criteria define the sample, the paper should show that the reported peaks are robust to the tolerance threshold and to the averaging procedure. A different threshold could alter the sample and the resulting azimuth distribution, affecting the central claim of non-random orientation and peak positions.
minor comments (7)
  1. [Abstract] The phrase 'without doubts' overstates certainty for a study that the author later calls 'preliminary' (Section 1). Soften the abstract to match the cautious language in the body.
  2. [§2] Typo: 'tecuques' should be 'techniques.' Also, 'Arce' appears in Section 4; should be 'Acre.'
  3. [§2] The estimated uncertainty of ±1° in azimuth is plausible but not justified by any inter-operator test. A short reproducibility check or reference to prior work would strengthen this.
  4. [§3] The histogram (Fig. 2) and curvigram (Fig. 3) would benefit from error bars or confidence bands, especially given the small counts in some bins.
  5. [§3] The use of a 3-Gaussian mixture is stated, but the number of components is a free parameter. Show sensitivity to the number of components, not only to the kernel type (Epanechnikov).
  6. [§1] Reference typo: 'Shahn 2012' should be 'Schaan 2012.' Also check the citation 'Kalliola, et al. 2024,2026' for consistency with the reference list.
  7. [§4] The suggestion that rhomboidal structures are ceremonial because no practical purpose is apparent is an argument from ignorance. Present this as a tentative observation, not as supporting evidence for the main conclusion.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: measured azimuths are compared to independently computable solar positions; self-citations are methodological only.

full rationale

The derivation chain is: extract azimuths of 324 rectangular earthworks from satellite imagery; fold each rectangle's two side bearings by selecting the side closer to due east; build kernel-density peaks; test against a uniform null; compare peak positions with the solar rising arc (66-114 deg) and with solstice/zenith values. Each ingredient is external to the conclusion: the solar arc, solstice azimuths, and calendar dates are computed from latitude and standard astronomy, and no parameter is fitted to make the peaks coincide with them. The folding rule, while described by a self-citation (Magli 2008), is a measurement convention; under an isotropic orientation null it maps a uniform axial distribution to a uniform distribution on 45-135, so it does not manufacture the observed peaks. The 68 deg peak is interpreted post hoc via the local agricultural calendar; that is an explanatory inference, not a quantity derived from the data that then reappears as a claim. The author's self-citations (Magli 2008, 2018, 2025) support method and general context, not the Amazonian result, and no uniqueness theorem or ansatz is imported from them. The paper also explicitly acknowledges 'this design may be due to practical reasons' and calls the results preliminary; this is a caveat about alternative explanations, not circular reasoning. The phrase 'borderline evidence (p ≈ 0.5)' appears internally inconsistent (p=0.5 is non-significant if literal), and the lack of a control sample for roads or terrain is a substantive weakness, but those are correctness/robustness concerns, not circularity. No equation in the paper is equivalent by construction to its inputs.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

The analysis itself uses standard statistics and external solar geometry; no invented entities. The listed free parameters and axioms are the places where the result could be tuned or confounded.

free parameters (4)
  • Internal-angle inclusion threshold = 2°
    Parallelograms whose internal angles deviate by more than 2° from 90° are excluded; this threshold is chosen by hand and changes the sample, but its influence on the peaks is not explored.
  • Number of Gaussian mixture components = 3
    The curvigram is presented as a 3-Gaussian mixture; the number of components is not selected by an information criterion and appears chosen to match the three visible peaks.
  • Kernel bandwidth = not stated
    The smoothing bandwidth of the Gaussian kernels is not reported, so the peak locations and significance values cannot be reproduced or checked for sensitivity.
  • Average latitude = 9°30' S
    A single latitude is used to compute the solar arc and solstice azimuths for all 324 sites; the paper states this is an average, without a sensitivity analysis over the actual latitude range (about 7°–11° S).
assumptions (5)
  • domain assumption Satellite-derived side azimuths represent original construction orientation within ±1°.
    The whole peak structure could be smeared or shifted by erosion, vegetation, and image interpretation; the paper assumes this uncertainty but does not propagate it into the significance tests.
  • domain assumption Uniform folded azimuth is the correct null distribution for random orientation.
    The global p-value of 1.5×10⁻¹⁵ is computed against uniformity; no alternative null reflecting terrain or practical constraints is considered (Section 2, Section 3).
  • domain assumption The side closest to due east is the meaningful orientation axis.
    Borrowed from Roman town orientation studies (Magli 2008); if geoglyph builders oriented the diagonals, the longer axis, or an entrance/road axis, the analysis could miss or distort the signal.
  • ad hoc to paper Modern Acre rainfall and double-cropping schedule apply to the ancient builders' calendar.
    The 68° peak is tied to end-of-May/mid-July dates using current rainfall (Rio Branco) and present-day corn harvests (Section 4), with no paleoclimate or archaeobotanical evidence.
  • domain assumption Flat-horizon solar arc is adequate; forest horizon effects are negligible.
    The paper computes solstice azimuths for a flat horizon and dismisses tree-horizon shifts as unlikely (Section 3), despite listing tree cover as an open question.

how reviews work

0 comments
Cite this review

Pith. "Pith review of The orientation of the Amazonian geoglyphs as a clue for their interpretation." pith.science (2026). https://pith.science/paper/ZZIHM3QX

@misc{pith2026260714201,
  author       = {Pith},
  title        = {Pith review of: The orientation of the Amazonian geoglyphs as a clue for their interpretation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ZZIHM3QX}},
  note         = {Machine review of arXiv:2607.14201}
}
read the original abstract

Amazonian earthworks, also called geoglyphs, are thousands of man-made earthen structures, mostly of geometrical shape, which progressively emerged from the tropical forest due to progressive deforestation. They were probably built between the fifth century BC and the end of the first millennium AD, but archaeological investigation on the culture of their builders is yet at the beginning. Nevertheless, a ceremonial rather than practical function seems likely, at least for those having a very regular shape. In the present paper, simple remote-sensing technique, combined with the methodological approach of modern Archaeoastronomy, are applied to study for the first time their orientation. The analysis takes in consideration virtually all known squared and rectangular structures for a total of 326 earthworks. The results show without doubts a non-random choice for their orientation and a clear interest of their builders for the annual cycle of the Sun.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

3 extracted references

  1. [1]

    D 2024 Geography of ancient geometric earthworks and their builders in southwestern Amazonia Acta Amaz

    Kalliola, R., Pärssinen, M., Ranzi, A., Seppa, I., & Barbosa, A. D 2024 Geography of ancient geometric earthworks and their builders in southwestern Amazonia Acta Amaz. 54 (4) • Oct-Dec 2024 Kalliola, R., Pärssinen, M., Ranzi, A., & Barbosa, A. D. (2026). Ancient Amazonian Earthwork Roads: Unveiling Ceremonial, Livelihood, and Networking Significances. La...

    2024

  2. [1095]

    Pärssinen, M., Balée, W., Ranzi, A., & Barbosa, A. (2020). The geoglyph sites of Acre, Brazil: 10 000-year-old land-use practices and climate change in Amazonia. Antiquity, 94(378), 1538-1556. Pärssinen, M. (2021). Tequinho Geoglyph Site and Early Polychrome Horizon 300 BC- AD 300/500 in the Brazilian State of Acre. Amazonica: revista de antropologia, 13(...

    2020

  3. [2012]

    Schaan, D., Pärssinen, M., Saunaluoma, S., Ranzi, A., Bueno, M., & Barbosa, A

    Sacred Geographies of Ancient Amazonia: Historical Ecology of Social Complexity Left Coast Press, Walnut Creek, 233p. Schaan, D., Pärssinen, M., Saunaluoma, S., Ranzi, A., Bueno, M., & Barbosa, A. (2012). New radiometric dates for precolumbian (2000–700 BP) earthworks in western Amazonia, Brazil. Journal of Field Archaeology, 37(2), 132-142. Saunaluoma, S...

    2012

Pith tools

Reviewed August 2, 2026 · model on record in the stance chip above.