Pith. sign in

REVIEW 4 major objections 6 minor 1 cited by

Structural Divergence of the Roman--Byzantine Trade Network, 0--1453\,CE: Persistent Homology, Topological Velocity, and Criticality Indicators of Imperial Collapse

T0 review · 4 major / 6 minor · reviewed 2026-07-11 · grok-4.5

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

Pith's one-line read Both the Western Roman collapse and the Byzantine endpoint land at the same topological entropy level H*≈0.524, while an East–West structural gap was already present at year 0 and commercial resilience could decouple from territorial contro

desk verdict Useful multi-layer TDA of the full Roman–Byzantine network with real new observables; the shared H*≈0.524 for 476 and 1453 is imposed by calibrating α0, not independently discovered. read the letter →

arxiv 2607.05695 v1 pith:CFERVGYO submitted 2026-07-06 physics.soc-ph econ.GNphysics.hist-phq-fin.EC

Jose de Jesus Bernal-Alvarado , David Delepine , Carlos Pinedo Guadarrama This is my paper · ORCID
classification physics.soc-phecon.GNphysics.hist-phq-fin.EC
keywords persistenthomologyRomantradenetworkByzantineresiliencetopologicalentropyWassersteinvelocitycollapsegeographic-economicdecouplingpercolationthreshold
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 maps the Roman–Byzantine trade system from 0 to 1453 CE as a time-evolving layered network and measures its cycle redundancy with persistent homology. It argues that the West was already topologically weaker than the East at the start of the Principate, that this gap widened for centuries before any formal political split, and that Byzantine survival rested on a commercial layer that could keep far more routing redundancy than the shrinking territorial layer. A single numerical threshold near 0.52 marks both the Western fall and the final Byzantine end, while the rate of change of the persistence diagrams picks out the Late Roman–Early Byzantine transition as the sharpest reorganisation in the whole record. A sympathetic reader cares because the method turns long-running debates about collapse versus continuity into measurable ratios between network layers, and because it offers a candidate early-warning signature for when an imperial route system loses spare topological capacity.

What carries the argument

β1 persistent entropy H of the trade network under a differential-friction filtration, split into geographic (H_geo) and economic (H_eco) layers whose ratio R_d = H_eco/H_geo measures decoupling; supplemented by inter-decade W2 Wasserstein velocity of persistence diagrams and a combined resilience index H_combined that mixes the two layers with a historically constrained capture factor.

What would settle it

Rebuild H_combined without fixing its free parameter to H*(1453)=0.524, or complete northern Gaul node coverage and recompute the East–West gap and terminal entropies; if the two collapse dates no longer land near the same H* or the congenital gap disappears, the threshold and asymmetry claims fail.

Watch

Extended reading notes

Core claim

With full western coverage, a baseline East–West entropy gap of about +2.22 units is present from 0 CE and grows at roughly +0.0033 per year; Byzantine geographic and economic layers then decouple, with the commercial-to-geographic ratio peaking near 48 around 620 CE. Independently, both the Western collapse in 476 CE and the Byzantine endpoint in 1453 CE occur at H*≈0.524, which the authors treat as a candidate topological percolation threshold for this system.

Load-bearing premise

The free parameter in the combined resilience index is set so that the index is forced to equal the same collapse level at 1453 that was observed for the West in 476, so the shared threshold is partly built in by calibration rather than found twice independently.

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 / 6 minor

Summary. The paper extends persistent-homology analysis of Mediterranean trade networks from 0–400 CE to the full Roman–Byzantine period (0–1453 CE) using 2,599 nodes and 4,503 trimodal edges. It reports five main results: (i) the prior H_t=0 western finding is a coverage artifact; with full western data a baseline East–West entropy gap of +2.22 already exists at 0 CE and grows at +3.3×10^{-3} yr^{-1}; (ii) top-degree sampling can reverse Phase-III slope signs (hub-selection artifact); (iii) Byzantine resilience decomposes into geographic and economic layers with peak decoupling R_d=H_eco/H_geo=47.7 at 620 CE; (iv) W_2 topological velocity peaks at the Late Roman–Early Byzantine transition (495 CE), and cross-network Wasserstein ratios jump 150–300× after the Chrysobull of 1082; (v) both the Western collapse (476) and Byzantine endpoint (1453) are said to occur at H*≈0.524, proposed as a candidate topological percolation threshold. Methods include differential friction, adaptive Vietoris–Rips filtration, Chow tests, a constructed H_combined index, and an Integrated Criticality Threshold.

Significance. If the non-circular results hold, the paper makes a genuine contribution to quantitative economic history and network science: it supplies a falsifiable, layer-resolved metric (R_d) that cleanly separates the Ward-Perkins and McCormick accounts, documents a congenital East–West structural asymmetry predating the Theodosian partition, and introduces practical methodological warnings (hub-selection artifact, minimum-coverage protocol) that apply beyond this case. The Wasserstein-velocity and cross-diagram analyses are model-free dynamical observables that strengthen the decoupling claim independently of entropy scalars. The dual-collapse H* claim, if it can be shown to be independently measured rather than imposed by calibration, would be a striking candidate criticality indicator; even without that claim the paper remains significant for layered TDA of historical systems.

major comments (4)
  1. §II.E.c, Eqs. (5)–(7) and Table III: the dual-collapse claim that both 476 CE and 1453 CE occur at H*≈0.524 is load-bearing in the abstract and conclusions, yet α0 is the sole free parameter and is fixed by the explicit constraint H_combined(1453)=H*=0.524, where H* is taken from the observed Western entropy at 476. The numerical coincidence for the Byzantine endpoint is therefore imposed by construction rather than independently measured. Pure H_geo already lies well below 0.524 for centuries while H_eco remains high; H_combined is engineered to hit the Western value at 1453. The paper should either (a) reframe H* as a calibration target and report the out-of-sample predictive content of the remaining historically fixed parameters (γ, λ, r_i) without claiming discovery of a shared threshold, or (b) define an independent Byzantine observable that crosses ~0.524 without using the Western
  2. §III (information-theoretic foundation of H*≈1/2): the two-bar argument yielding H*∈[0.500,0.572] is presented as supporting the observed 0.524, but it rests on the geometric ansatz ℓ_i∝D_i and an assumed r_c≈0.25–0.30 for Roman/Byzantine scales, plus a noise-floor cut p_2≥0.20. The manuscript itself notes that real terminal diagrams have 2–4 bars and that weighted LCP filtrations introduce non-linear corrections of order σ_cost². This interval is therefore a plausible range, not an independent derivation of the specific value 0.524. The section should be clearly labelled as a consistency check / order-of-magnitude argument, not as a first-principles derivation that removes the calibration circularity of H_combined.
  3. §II.A and footnote [26]: western coverage remains incomplete for Gallia (only 9 nodes vs 443 Hispania / 141 Africa Occidentalis). The authors correctly flag that ΔH_0=+2.22 should be treated as an upper bound pending fuller Gallia data. Because the congenital East–West asymmetry is a central claim (result i), the paper should either augment Gallia from DARE/Barrington or provide a quantitative sensitivity analysis (e.g., synthetic Gallia nodes at plausible densities) showing that the sign and approximate magnitude of the gap, and the Phase-II western degradation, survive reasonable coverage corrections. Without that, the precise gap value and growth rate remain provisional.
  4. Bootstrap and Monte Carlo uncertainty: Byzantine bootstrap CIs are wide (mean 95% widths ~0.715 for H_geo and ~0.503 for H_eco at N_core=400 FPS subsample), while the main analysis uses N_core=1,600. The H_combined Monte Carlo (Fig. 1) places historical 1453 at the 100th percentile of the crossing distribution (mean 1313±33). This is consistent with a late, forced hit under the α0 calibration but weakens any claim that H_combined independently forecasts 1453. Report full-N bootstrap CIs for the main entropy series where feasible, and present the 1453 hit as a calibration outcome rather than a prediction unless an unfitted forecast is shown.
minor comments (6)
  1. Abstract and §VII: the five-result list is clear, but the dual-collapse H* sentence should be reworded once the calibration issue is resolved so that readers are not led to treat 0.524 as an independently discovered constant.
  2. Table V and §II.G: the hub-selection artifact is valuable and well documented; consider adding a short recommended sampling protocol (e.g., stratified by longitude or province) so that others can apply the minimum-coverage check without re-deriving N_min.
  3. Roman vs Byzantine W_2 units: footnote [23] correctly notes that absolute magnitudes are not comparable across periods. State this once in the main text near Tables VII and XI to avoid misreading of velocity rankings across eras.
  4. Notation: H*, H_t, H_geo, H_eco, H_combined, and R_d are introduced in several places; a single notation table early in §II would help. Also fix occasional typographic inconsistencies (e.g., H t=0 vs H_t=0, ce vs CE).
  5. Fig. 4 / Fig. 6: ensure axis labels and the H*=0.524 line remain legible in grayscale; the dual-axis R_d panel is dense.
  6. References: companion paper [1] is cited as arXiv:2605.27200; ensure the final version points to a stable identifier and that ORBIS/Pleiades/DARMC/CHRR access dates or versions are recorded for reproducibility.

Circularity Check

2 steps flagged · score 7.0 of 10

H* dual-collapse claim is forced by calibrating the sole free parameter α0 so H_combined(1453) equals the observed Western H(476) by construction

  1. fitted input called prediction [§II.E.c, Eqs. (5)–(7) and Table III]
    "the single calibration constraint is Hcombined(1453ce) = H ∗ = 0.524, where H ∗ is the observed topological level of the western network at 476ce. Only α0 is a free parameter, fixed by requiring that H combined(1453ce) = H ∗ = 0.524. This fixes the value of α 0 ≈1.2031. … α 0 is the sole calibrated parameter, fixed by the H ∗ constraint."

    H* is taken from the observed Western entropy at collapse; α0 is then solved so that the constructed H_combined trajectory equals that same numerical value exactly at 1453 CE. The subsequent claim that the Byzantine endpoint also occurs at H*≈0.524 is therefore true by the calibration constraint rather than discovered from an independent observable.

  2. self definitional [Abstract; §VII Conclusions; Fig. 1 caption]
    "Both the Western collapse (476ce) and the Byzantine endpoint (1453ce) occur at H∗≈0.524, interpreted as a candidate topological percolation threshold. … Both the Western Roman collapse (476,ce) and the Byzantine collapse (1453,ce) occur at H ∗ = 0.524±0.031 … The horizontal dotted line marks H ∗ = 0.524. … the purple star marks the historical date (1453ce) at the 100th percentile of the crossing distribution."

    The dual-occurrence statement treats the calibrated H_combined(1453) as an independent observation of the same threshold that was measured for the West. Because H_combined is defined to equal H* at 1453, the statement reduces to ‘the Western value equals the value we forced the Byzantine composite to equal’.

full rationale

The paper's central criticality claim—that both the Western Roman collapse (476 CE) and Byzantine endpoint (1453 CE) occur at the same H*≈0.524, a candidate topological percolation threshold—is only partially independent. H* is first observed as the Western network entropy at 476 CE. H_combined is then explicitly constructed (Eqs. 5–7) with historically fixed parameters except for the single free parameter α0, which is fixed by the calibration constraint H_combined(1453 CE)=H*=0.524. Table III and the surrounding text state this openly. Consequently the numerical coincidence for the Byzantine endpoint is definitional for the composite index rather than an independent measurement; pure H_geo already sits far below 0.524 for centuries while H_eco remains high. The two-bar information-theoretic interval [0.500,0.572] in §III supplies a plausible external range that 0.524 falls inside, and the East–West gap, Rd, Wasserstein velocity and ICT results are independent of this calibration, so the circularity is partial (score 7) rather than total. Monte-Carlo bands around the forced trajectory do not remove the construction.

Assumptions & free parameters 6 free parameters · 4 assumptions · 4 invented entities

The central threshold and combined-resilience claims rest on one fitted scalar (α0), several historically motivated but free multipliers, the identification of H* with the Western 476 value, and the modeling choice that commercial routes persist after territorial loss. The baseline asymmetry and raw H_geo/H_eco series rest on fewer free parameters.

free parameters (6)
  • α0 (commercial-capture efficiency prefactor) = ≈1.2031
    Sole free parameter fixed by the constraint H_combined(1453)=0.524; value ≈1.2031.
  • γ (territorial-loss exponent in w(t)) = 0.764
    Set to 0.764 from Hendy’s fiscal-capacity estimate; controls how fast geographic weight declines.
  • λ (base commercial-capture decay) = 5.0e-4 yr^{-1}
    5.0×10^{-4} yr^{-1} taken from Laiou’s long-run fiscal contraction rate.
  • shock multipliers r_i (Arab, Macedonian, Chrysobull, Palaeologan) = 0.59 / 1.36 / 0.44 / 0.55
    Four discrete multipliers (0.59, 1.36, 0.44, 0.55) taken from secondary literature and applied to α(t).
  • H* collapse threshold = 0.524
    Taken as the observed Western entropy at 476 CE (≈0.524) and then imposed on the Byzantine combined series.
  • road-cost calibration factor = 8.90×
    8.90× from log-log regression on 42 ORBIS–road pairs; 95 % CI 7.31–10.84.
assumptions (4)
  • domain assumption Commercial routes continue to exist (nodes retained, only costs change) after loss of administrative control; this defines the economic layer G_eco.
    Stated in §II.E; without it H_eco cannot remain high after territorial loss and Rd is undefined.
  • ad hoc to paper Bar lifetimes in the Vietoris–Rips filtration are proportional to geographic cycle diameter, so the two-bar entropy H2(rc) with rc≈0.25–0.30 yields H*∈[0.50,0.57].
    §III; the proportionality and the specific rc values are modeling choices, not derived from the data.
  • domain assumption Adaptive filtration threshold δt = 90th percentile of pairwise distances is an appropriate scale for historical route networks.
    Inherited from the companion paper and used throughout the pipeline.
  • standard math Chow structural-break tests on the entropy time series correctly identify historically meaningful phase boundaries.
    Standard econometric tool applied to the topological series; significance levels are reported.
invented entities (4)
  • H_combined (combined resilience index)
    purpose: Scalar that mixes geographic and economic entropy with time-varying territorial weight and fiscal-capture efficiency so that a single threshold can be compared to both collapses.
    Defined by Eqs. 5–7; calibrated to hit H* at 1453; no independent measurement outside the model.
  • Integrated Criticality Threshold (ICT)
    purpose: Average of normalised susceptibility, correlation length and Wasserstein velocity intended as an early-warning indicator of topological phase transition.
    Eq. 1 / Eq. 15; composite of three PD-derived quantities; behaviour near 1 is interpreted as criticality by analogy.
  • Topological velocity Ẇ2 independent evidence
    purpose: Decade-to-decade W2 distance between persistence diagrams, used to rank historical events by structural reorganisation speed.
    Standard Wasserstein metric applied to consecutive PDs; the ranking itself is new for this corpus.
  • Hub-selection artifact independent evidence
    purpose: Name for the non-monotonic N-convergence of Chow statistics under top-degree sampling in geographically skewed networks.
    Documented in Table V; methodological warning rather than a physical entity.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Structural Divergence of the Roman--Byzantine Trade Network, 0--1453\,CE: Persistent Homology, Topological Velocity, and Criticality Indicators of Imperial Collapse." pith.science (2026). https://pith.science/paper/CFERVGYO

@misc{pith2026260705695,
  author       = {Pith},
  title        = {Pith review of: Structural Divergence of the Roman--Byzantine Trade Network, 0--1453\,CE: Persistent Homology, Topological Velocity, and Criticality Indicators of Imperial Collapse},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/CFERVGYO}},
  note         = {Machine review of arXiv:2607.05695}
}
abstract

We extend the persistent homology analysis of~\paperone{} to the full Roman--Byzantine trade network (0--1453\,\textsc{ce}), using 2{,}599 nodes and 4{,}503 trimodal edges calibrated against the \textsc{orbis} Geospatial Network Model. Five results are reported. % (i)~The $H_t{=}0$ western sub-network result of~\paperone{} is a data-coverage artifact: with full western representation ($N_{\rm west}=987$, $\beta_1\approx52$ cycles per decade) a baseline East--West entropy gap of $+2.22$ units is present from 0\,\textsc{ce} and grows at $+3.3\times10^{-3}$\,yr$^{-1}$, predating the Theodosian partition by four centuries. % (ii)~A \emph{hub-selection artifact} in degree-heterogeneous networks can reverse the sign of the inferred Phase~III slope, requiring full-coverage or stratified sampling for reliable structural-break detection. % (iii)~Decomposing Byzantine resilience into geographic ($H_{\rm geo}$) and economic ($H_{\rm eco}$) components reveals a peak decoupling ratio $R_d = H_{\rm eco}/H_{\rm geo} = 47.7$ at 620\,\textsc{ce}, falling to 13.9 at 640\,\textsc{ce}, quantifying the McCormick--Ward-Perkins historiographical debate as a contrast between two network layers operating on different timescales. % (iv)~The inter-decade $W_2$ Wasserstein velocity identifies the Late Roman--Early Byzantine transition (495\,\textsc{ce}) as the highest topological-velocity event of the 1,453-year record; the cross-network Wasserstein ratio increases by $150$--$300\times$ after the Chrysobull of 1082\,\textsc{ce}, providing an independent diagram-space analogue of $R_d$. Both the Western collapse (476\,\textsc{ce}) and the Byzantine endpoint (1453\,\textsc{ce}) occur at $H^{\ast}\approx0.524$, interpreted as a candidate topological percolation threshold.

Figures

Figures reproduced from arXiv: 2607.05695 by the authors.

Figure 1
Figure 1. FIG. 1. Monte Carlo uncertainty on the combined resilience index [PITH_FULL_IMAGE:figures/full_fig_p007_1.png] view at source ↗
Figure 2
Figure 2. FIG. 2. (a) [PITH_FULL_IMAGE:figures/full_fig_p013_2.png] view at source ↗
Figure 3
Figure 3. FIG. 3. Decadal coin-hoard counts by region, 200–320 [PITH_FULL_IMAGE:figures/full_fig_p015_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: FIG. 4. Geographic entropy [PITH_FULL_IMAGE:figures/full_fig_p019_4.png]
Figure 5
Figure 5. Figure 5: FIG. 5. Economic entropy [PITH_FULL_IMAGE:figures/full_fig_p020_5.png]
Figure 6
Figure 6. Figure 6: FIG. 6 [PITH_FULL_IMAGE:figures/full_fig_p023_6.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Topological Signatures of Imperial Collapse and Fragmentation: Administrative Dissolution, Territorial Reorganization and Early-Warning Observables in the Han Dynasty Network (206~BCE\,--\,220~CE)

    physics.soc-ph 2026-07 conditional novelty 5.0 of 10

    Han administrative persistent entropy collapses to zero at 220 CE while geographic entropy rises to 3.065, with Wasserstein and ICT early warnings 45–50 years prior and Three Kingdoms as β1 cycles at 190 CE.

Reference graph

Works this paper leans on

32 extracted references · 32 canonical work pages · cited by 1 Pith paper

  1. [26]

    Kennedy,The Prophet and the Age of the Caliphates: The Islamic Near East from the Sixth to the Eleventh Century (Longman, London and New York, 1986)

    H. Kennedy,The Prophet and the Age of the Caliphates: The Islamic Near East from the Sixth to the Eleventh Century (Longman, London and New York, 1986)

    work page 1986

  2. [1]

    Wasserstein velocity From theW 2 Wasserstein distance between consecutive persistence diagrams, the topological velocity ˙W2(t)≡ W2(Dt,D t+10)/10 [15, 16], provides a model-free measure of the rate at which the topological structure of the network is changing. Unlike the persistent entropyH(t), which is a scalar summary of the diagram at a single decade, ...

    work page

  3. [2]

    Integrated Criticality Threshold: Roman period Table VIII reports the phase-meanictfor all four Roman-period PD sub-datasets. The Full, Geo, and Trade PD sub-datasets sustainict≈0.30–0.46 throughout Phases I–III, indicating that the large, densely connected Roman network was never far from a critical state during the study period. Theictof the Full networ...

    work page

  4. [3]

    The values are in normalised cost units and are not directly comparable in absolute magnitude to the Roman-period values, which useorbisdenarius units

    Wasserstein velocity Table XI reports the smoothed Wasserstein velocity ˙W2,sm for the Geo PD sub-dataset at the documented Byzantine break events. The values are in normalised cost units and are not directly comparable in absolute magnitude to the Roman-period values, which useorbisdenarius units. However, the relative ordering of events is internally co...

    work page

  5. [4]

    The administrative, commercial, and geographic cycle structures deform at the same rate

    Cross-network Wasserstein divergence and the Chrysobull decoupling The cross-network Wasserstein distanceW cross between the Admin and Geo persistence diagrams (the diagram- space analogue of the decoupling ratioR d =H eco/Hgeo) shows a dramatic and abrupt transition at 641–700ceand again at 1080ce: •Before 640ce:the ratioW cross(Admin,Geo)/W cross(Admin,...

    work page

  6. [5]

    Three patterns emerge

    Integrated Criticality Threshold: Byzantine period Table XIII reports theictfor all four Byzantine networks by historical phase. Three patterns emerge. a. Admin PD sub-dataset: criticality collapses monotonically.The Admin PD sub-dataset (the Byzantine-period administratively controlled node set; same spatial extent asGgeo t but not used forH geo entropy ...

    work page

  7. [6]

    J. d. J. Bernal-Alvarado, D. Delepine, and C. Pinedo Guadarrama, arXiv preprint2605.27200(2026), arXiv:2605.27200

    work page Pith review arXiv 2026

  8. [7]

    Meeks and R

    E. Meeks and R. Mostern, ORBIS: The Stanford geospatial network model of the Roman world, Stanford Humanities + Design Lab (2013)

    work page 2013

Show all 32 references
  1. [8]

    Scheidel and E

    W. Scheidel and E. Meeks,ORBIS: The Stanford Geospatial Network Model of the Roman World, version 2.0, Tech. Rep. (Stanford University, 2014)

    2014

  2. [9]

    Scheidel, I

    W. Scheidel, I. Morris, and R. P. Saller, eds.,The Cambridge Economic History of the Greco-Roman World(Cambridge University Press, Cambridge, 2007)

    2007

  3. [10]

    R. J. A. Talbert, R. S. Bagnall, T. Elliott, S. Gillies,et al., Pleiades: A community-built gazetteer and graph of ancient places, Pleiades Project (2024)

    2024

  4. [11]

    McCormicket al., Digital atlas of Roman and Medieval civilizations (DARMC), Harvard University, Mapping Past Societies (2013)

    M. McCormicket al., Digital atlas of Roman and Medieval civilizations (DARMC), Harvard University, Mapping Past Societies (2013)

    2013

  5. [12]

    ˚Ahlfeldt, Digital atlas of the Roman empire (DARE), Digital Atlas of the Roman Empire (2013)

    J. ˚Ahlfeldt, Digital atlas of the Roman empire (DARE), Digital Atlas of the Roman Empire (2013)

    2013

  6. [13]

    Wickham,Framing the Early Middle Ages: Europe and the Mediterranean, 400–800(Oxford University Press, Oxford, 2005)

    C. Wickham,Framing the Early Middle Ages: Europe and the Mediterranean, 400–800(Oxford University Press, Oxford, 2005)

    2005

  7. [14]

    Heather,The Fall of the Roman Empire: A New History of Rome and the Barbarians(Pan Books, London, 2006)

    P. Heather,The Fall of the Roman Empire: A New History of Rome and the Barbarians(Pan Books, London, 2006)

    2006

  8. [15]

    Ward-Perkins,The Fall of Rome and the End of Civilization(Oxford University Press, Oxford, 2005)

    B. Ward-Perkins,The Fall of Rome and the End of Civilization(Oxford University Press, Oxford, 2005)

    2005

  9. [16]

    McCormick,Origins of the European Economy: Communications and Commerce, A.D

    M. McCormick,Origins of the European Economy: Communications and Commerce, A.D. 300–900(Cambridge University Press, Cambridge, 2001)

    2001

  10. [17]

    Harper,The Fate of Rome: Climate, Disease, and the End of an Empire(Princeton University Press, Princeton, 2017)

    K. Harper,The Fate of Rome: Climate, Disease, and the End of an Empire(Princeton University Press, Princeton, 2017)

    2017

  11. [18]

    G. C. Chow, Econometrica28, 591 (1960)

    1960

  12. [19]

    J. R. McConnell, A. I. Wilson, A. Stohl, M. M. Arienzo, N. J. Chellman, S. Eckhardt, E. M. Thompson, A. M. Pollard, and J. P. Steffensen, Proceedings of the National Academy of Sciences of the United States of America115, 5726 (2018)

    2018

  13. [20]

    Edelsbrunner, D

    H. Edelsbrunner, D. Letscher, and A. Zomorodian, Discrete & Computational Geometry28, 511 (2002)

    2002

  14. [21]

    Kerber, D

    M. Kerber, D. Morozov, and A. Nigmetov, inJournal of Experimental Algorithmics, Vol. 22 (2017) pp. 1.4:1–1.4:20

    2017

  15. [22]

    Coin Hoards of the Roman Republic, Coin hoards of the roman republic database,https://coinhoards.org(2025)

    2025

  16. [23]

    Naismith, R

    R. Naismith, R. Bland, M. Allen, and M. Lewis, eds.,Coin Hoards and Hoarding in the Roman World(Oxford University Press, Oxford, 2022)

    2022

  17. [24]

    M. F. Hendy,Studies in the Byzantine Monetary Economy, c. 300–1450(Cambridge University Press, Cambridge, 1985)

    1985

  18. [25]

    A. E. Laiou, ed.,The Economic History of Byzantium: From the Seventh Through the Fifteenth Century(Dumbarton Oaks Research Library and Collection, Washington, D.C., 2002)

    2002

  19. [27]

    Rucco, F

    M. Rucco, F. Castiglione, E. Merelli, and M. Pettini, Entropy18, 117 (2016)

    2016

  20. [28]

    Their absolute magnitudes are therefore not directly comparable

    The Roman-period WassersteinW 2 values are expressed inorbisdenarius cost units, while the Byzantine-period values are in normalised units from a separate pipeline. Their absolute magnitudes are therefore not directly comparable. All inter- period statements in this paper are ...

  21. [29]

    B”untgen, V

    U. B”untgen, V. S. Myglan, F. C. Ljungqvist, M. McCormick, N. Di Cosmo, M. Sigl, J. Jungclaus, S. Wagner, P. J. Krusic, J. Esper, J. O. Kaplan, M. A. C. de Vaan, J. Luterbacher, L. Wacker, W. Tegel, and A. V. Kirdyanov, Nature Geoscience 9, 231 (2016)

    2016

  22. [30]

    Mordechai, M

    L. Mordechai, M. Eisenberg, T. P. Newfield, A. Izdebski, J. E. Kay, and H. Poinar, Proceedings of the National Academy of Sciences116, 25546 (2019)

    2019

  23. [31]

    443 for Hispania and 141 for Africa Occidentalis), reflecting the incompleteness of the Pleiades coverage for northern Gaul

    The western sub-network currently contains only 9 Gallia nodes (vs. 443 for Hispania and 141 for Africa Occidentalis), reflecting the incompleteness of the Pleiades coverage for northern Gaul. Augmenting the Gallia coverage from the full DARE dataset and from the Barrington At...

  24. [32]

    R. J. A. Talbert, ed.,Barrington Atlas of the Greek and Roman World(Princeton University Press, Princeton, 2000)

    2000

Pith tools

Reviewed July 11, 2026 · model on record in the stance chip above.