What this volume got wrong

contested — 42

Not hedges. Each of these is a location: the row says where the sources part company and what it would take to settle it.

1. Puha — power, and where it livescontinuous; recorded in ethnography from the 1930sch. I

   Recorded in Park, Shamanism in Western North America (1938); Kelly, Ethnography of the Surprise Valley Paiute (1932); Stewart’s culture-element distributions (1941); Miller on puha (1983); Handbook of North American Indians vol. 11 (1986). ⚠ All of it is reservation-period recollection, set down sixty to seventy years after the wars — no source describes these practices from the inside at the time. And some of this material is held by the tribes as restricted knowledge, which published ethnography does not settle. Sacred site locations are deliberately absent from the map on this page and should stay absent.

2. Why “witchcraft” is the wrong wordcontinuousch. I

   As the puha row: reservation-period ethnography, principally Park (1938). The category error being corrected here is the observers’, not the sources’.

3. The antelope charmercontinuousch. I

   Recorded in the same 1930s ethnography. Steward’s “gastronomic orientation” is the standard characterisation of his framing; Park’s work on shamanism is usually named as the exception to it. ⚠ Same sourcing limitation as the rows above.

4. The oldest art on the continent, on a dead lakebeddated to roughly 10,500–14,800 years before presentch. I

   🔴 RESTRICTED. No location is given here at any precision and none should be published — rock art sites are withheld to prevent vandalism. The dating range is wide because it rests on carbonate-layer chronology tied to lake-level history rather than on the carvings themselves; cite the range, never a single figure.

5. Two water countries that do not touchpermanentch. I

   Four terminal northern rivers with no interbasin connection to the Colorado system; return-flow credits on the southern allocation. ⚠ Verify the current Colorado allocation figure before publishing anything numeric — shortage-tier reductions have moved it repeatedly since 2021, so any number here dates fast.

6. The free resource here is the night, not the windthe vernacular, and what still worksch. I

   Diurnal swing of 30–40°F and 8–10% RH; evaporative cooling efficiency and nocturnal radiative cooling both depend directly on low humidity and clear skies. TWO TECHNICAL POINTS worth more than the rest: ⚠ insulate heavily and put the MASS INSIDE the insulation — mass outboard of insulation is a liability in a cold winter, and that is exactly where naive adobe fails at this latitude. And Reno has a resource almost no American city has: the Moana field in the south-west of the city runs DIRECT-USE geothermal — hot water out of the ground into radiators, no electricity in between — which addresses the heating-dominated half that Egyptian design simply ignores. 🔴 TWO CONSTRAINTS EGYPT NEVER HAD, and the first is the bigger change from historic practice: SMOKE SEASON breaks any strategy built on opening the house at night, so a serious design here needs a sealed mode with filtered mechanical ventilation and heat recovery — the passive scheme must have a closed fallback. And WATER LAW: Nevada is a prior-appropriation state, so rain falling on a roof is arguably already appropriated downstream. Limited residential rooftop capture has been loosened in recent years but check the current statute, and check the separate greywater rules, before designing around either. Which is the joke at the end of this volume: you can manipulate heat, light and air on your own property with nothing but geometry and no one’s permission. The one element you cannot design around is the one the desert never had.

7. A road that gets worse as you use itevery season, worst 1849–1852ch. IV

   The Forty Mile Desert crossing and its carcass counts come from 1850 Army and emigrant tallies, which are estimates over a chaotic season rather than a census — the figures run into the thousands and are not precise. The degradation mechanism itself is uncontested and appears in the diaries directly.

8. The Donner Party and the Hastings Cutofftrapped late October 1846; last rescue April 1847ch. IV

   Usually given as 87 in the party with 48 surviving, but the count depends on where you start it — members joined and left en route. Cite it as approximate or not at all.

9. The Knockers, and why belief is infrastructureCornish miners from the 1820s; belief active into the 1930s; a petition in 1956ch. VII

   Cornish bucca / Welsh bwca carried by Cousin Jack miners; Billie Williams’s figures on the Comstock; the 1956 petition to release the knockers on a mine sealing. New Yellow Jacket inflow at the 3,080-ft level, 1880. ⚠ The 1956 petition and the Williams detail come through mining folklore collections rather than company records — good stories are exactly what such collections preserve best, so treat the specifics as reported.

10. How violent it actually wasmeasured across the 1860s–1880sch. VII

    Dykstra, The Cattle Towns (1968), 45 homicides across five towns 1870–85 — ⚠ often quoted as proving the West was peaceful, which misreads it: 45 deaths in populations that small is about 33 per 100,000 a year, high by any modern standard. Dykstra’s claim was that the reputation exceeded the facts, not that the facts were mild. McGrath, Gunfighters, Highwaymen, and Vigilantes: Bodie 31 homicides 1877–82, population 2,712; rates of 116 (Bodie) and 64 (Aurora). Roth, American Homicide, on legitimacy and fellow-feeling. Re-derive any rate before printing it — denominators in transient camps are the weakest part of all this work.

11. Where the ghosts actually come from1850s–1870sch. VII

    American Spiritualism’s peak 1850s–70s is uncontested; mediums working in Virginia City are reported in the local record. The Chinese exhumation-and-repatriation practice is well documented across the western camps. The Ghost Dance connection is drawn here as reading, not as a claim that the movements share a cause — see the accounts panel.

12. The Pyramid Lake War12 May 1860; second engagement June 1860ch. VII

    The Williams Station trigger — the two girls held beneath the cabin — is well attested and is carried by the Pyramid Lake Paiute Tribe’s own account and by modern local histories; contemporary settler newspapers omitted it. ⚠ The details of the girls’ treatment and the exact sequence come to us largely through Paiute oral testimony recorded later and through hostile settler sources, which is precisely the divergence the accounts panel below exists to hold. Three rows in the War Atlas: the war and both engagements.

13. The wars nobody namesOwens Valley 1862–63; the Snake War 1864–68ch. VII

    The 1,762 figure comes from modern compilation of scattered army returns rather than a contemporaneous count, and casualty totals for this war should always carry that caveat. ⚠ CALIBRATE THE SUPERLATIVE: “bloodiest conflict of the trans-Mississippi West” is a comparison INSIDE a theatre and does not survive being carried out of it — the entire trans-Mississippi West produced fewer deaths in thirty years than the Somme produced in a fortnight. This war’s claim on attention is that it is enormous for its region and almost unrecorded, not that it is large by world standards. ✅ AND THE ATLAS CLAIM IS NOW MEASURED RATHER THAN ASSERTED: an audit found the Snake War holds exactly one row against thirteen for the Wars of the Roses — a real EVENT deficit. But on a rows-per-casualty basis the bias runs the other way by roughly two orders of magnitude, because one Waterloo row absorbs 55,000 dead and one Somme row absorbs a million. The atlas indexes on NAMES, and a name is cheap in a small war and expensive in a large one. Both distortions are real and they run in opposite directions. See lib/nevada-atlas-audit.ts.

14. Nevada’s most important water is in another statelogged 1870s–80s; the outlet dammed 1913; TRPA compact 1969ch. VII

    Elevation ~6,225 ft, max depth ~1,645 ft, ~191 sq mi, ~72 mi shoreline; sole outlet the Truckee at Tahoe City, dammed 1913, top ~6 ft as storage; Truckee River Operating Agreement governs releases. Tahoe Regional Planning Agency created by bi-state compact in 1969 and ratified by Congress — a two-state land-use authority with real power, and perpetually contested. ⚠ Clarity figures move annually; take them from the current UC Davis State of the Lake report rather than from this row.

15. Virginia Citypeak mid-1870sch. VII

    Both halves need care. The 40,000 figure belongs to the whole Comstock district at its mid-1870s peak, not to Virginia City in 1864 — the city was materially smaller then, and the state census of 1875 put it near 20,000. And it was never the largest in the country; the usual claim is largest between Chicago and San Francisco, which is itself soft. Mining-camp counts are unreliable in both directions: transient populations are undercounted and boosters inflate.

16. Battle Born31 October 1864ch. VII

    Admission by presidential proclamation 31 October 1864; 36th state; capital Carson City. The telegraphed constitution is documented as the longest and most expensive telegram of its day. ⚠ But the re-election half does not survive the arithmetic: Lincoln took 212 electoral votes to McClellan’s 21, and Nevada’s three were not needed and were never close to being needed. The motives that hold up are the congressional ones — Republican margins in the House and Senate for the Thirteenth Amendment and for Reconstruction. The election story is the one everyone tells; it is the weakest version of a true claim.

17. Built the crossing, then voted atrailroad labour from 1868; the exclusion referendum 1879ch. VIII

    Chinese labour on the Central Pacific across Nevada from 1868; the 1879 statewide exclusion referendum; the practice of exhumation and repatriation of remains, documented across the western camps. ⚠ This row is thinner than it should be and is a known gap rather than a finished account — population figures, the Virginia City quarter, and the anti-Chinese expulsions of the 1880s all need proper sourcing before this stands as more than a marker.

18. Lake Street — the boardinghouse as an institutionat least ten boardinghouses in the 1920s–30s; the fire 15 August 1948ch. VIII

    Santa Fe building 1913, hotel under several names including the Colombo; Etcheberry lease 1931, Marie Etcheberry with Martin Esain after 1943; Lake Street fire 15 August 1948, five dead; rebuilt 1949. Louis’ Basque Corner opened 1967 in the 1906 Lincoln Hotel building — and was never a herders’ boardinghouse, which is a distinction routinely lost. ⚠ The Santa Fe’s restaurant closed in July 2020, reopened October 2023 as a pintxos bar, and a further closure was announced in December 2024 — that address has changed hands repeatedly, so check what is open before sending anyone. The sign has been preserved.

19. The archive carved into treeslate 1800s to the 1980sch. VIII

    ~25,000 documented, with estimates of a comparable number unrecorded; 70–80% of northern Nevada’s believed already lost. The Jon Bilbao Basque Library at UNR holds the Earl Collection — 130 rubbings taken from the 1970s — and the Mallea-Olaetxe Lertxunmarrak Collection. ⚠ Counts and loss estimates are researcher estimates over an incompletely surveyed landscape, not a census. Ranges are named here; specific carving locations are not, for the same vandalism reason that keeps rock art off this page’s map.

20. The Belmont mine fire, and the hotel that sells a different death23 February 1911ch. IX

    Fire 23 February 1911 at the 1,166-ft level; seventeen dead; Murphy’s two successful trips and his death on the third; concealment of the toll, an inquest of about seventy witnesses, no charges. ⚠ Corrections to the brief that raised this: the fire was at 1,166 feet, not five hundred; the superintendent is named in the record as T. F. M. Fitzgerald but the specific claim that he ordered men to work into a known fire is NOT confirmed by the sources reached here — what is documented is “a series of miscalculations and mistakes by senior officials.” Do not assert the order. The death toll is also given as fourteen in some tellings; pin it to the coroner’s record before quoting. On the ghost: no documented death underlies the Lady in Red — the method, floor and rooms all vary between tellings and no body, inquest or newspaper has been tied to her, which is the signature of folklore rather than of a killing. The hotel’s Wyatt Earp bar names a man who was in Tonopah around 1902 and never in a building that opened in 1907.

21. The Tonopah plagueusually told as 1902; the mortality record points to 1905ch. IX

    ⚠ Genuinely unresolved and carried here as a warning, not a fact. The ~300 figure is the cemetery’s total interments, not the outbreak’s toll — those are routinely conflated. And the State Board of Health record points to 56 deaths between January and April 1905 from pneumonia, attributed to poor sanitation near the slaughterhouses, rather than to a mysterious 1902 event. Cemetery closed April 1911. Resolve against the Board of Health returns before using any figure.

22. Fallon, Fernley, and the water taken from the lakeAct signed 1902; Derby Dam and the canal complete 1905ch. IX

    ✅ CORRECTED: this row previously called him Senator Newlands in 1902. He sat in the House from 1893 and entered the Senate on 4 March 1903 — after the Reclamation Act was signed on 17 June 1902. The act that carries his name was carried by a congressman. Reclamation Act 1902; project authorised 14 March 1903; Derby Dam begun October 1903 and completed 20 May 1905; canal ~31 miles; about 57,000 acres served. ⚠ The Newlands Project is very often called the first federal reclamation project and the sources split — some say first authorised, others place it among the original five. Say “one of the first,” or name the specific claim you mean.

23. Seniority concentrates, and juniors fragmentcomputed from the NDWR Newlands water-righted areas layer, August 2026ch. IX

    Computed from the published NDWR layer: 1,082 parcels, 28,889 acres. ⚠ THE ACREAGE DOES NOT RECONCILE with the ~57,000 acres the Newlands Project is usually said to serve, and the honest position is that I do not know why. The layer may cover one division rather than both, or one class of right rather than all. Do not report this as “the project halved” — resolve what the layer actually covers against Reclamation's own project figures first. That discrepancy is the most useful thing in this row, because it is checkable.

24. The town that sued to keep the leakbreach 5 January 2008; settlement approved 2016; the lining fight ongoingch. IX

    Breach 5 January 2008, ~590 homes, ~3,500 evacuated; prior 1996 rupture; rodent burrowing to ~25 ft; $18.1m class settlement approved 2016 covering about 1,200 people, financed by a TCID members’ vote to sell water rights to TMWA; the canal-lining litigation through the Ninth Circuit and the Nevada Supreme Court. ⚠ The lining case is live and its posture moves — verify status before citing an outcome. Fernley incorporated 2001, is around 20,000 people, has had an Amazon fulfilment centre since the late 1990s, and its Victory Logistics District drew a Microsoft land purchase reported at $70.5m in 2025: a town founded on irrigated alfalfa now running on warehouses and data centres, still depending legally and physically on a dirt ditch dug in 1905.

25. Lake Mead, and what it is nowfilled 1935–37; reading below dated July 2026ch. IX

    ⚠ “Largest reservoir in the United States” is true of its CAPACITY and badly misleading about its CONTENTS. As of July 2026 Lake Mead was about 25% full and had fallen to its lowest level since it was first filled in 1937; Lake Powell was near 22%. Never state the superlative without the date and the fill — the headline is a design specification, not a measurement, and on this page the distinction is the whole method.

26. The Moulin Rouge AgreementMoulin Rouge opened 1955; the agreement 26 March 1960ch. XI

    The agreement, its date and its participants are well documented. ⚠ The epithet “Mississippi of the West” is widely repeated of Las Vegas in this period but I could not source it to a contemporary use in this pass — treat it as reported, not established.

27. Twain and Steinbeck — the same road, told twiceTwain west 1861, Roughing It 1872; Steinbeck at Tahoe 1928, The Grapes of Wrath 1939ch. XIII

    Clemens in the Esmeralda and Humboldt districts 1861–62; Territorial Enterprise from February 1863; Roughing It 1872. Steinbeck met Carol Henning at Lake Tahoe, June 1928, while working at a fish hatchery; Gwyn Steinbeck obtained a Reno decree in October 1948. ⚠ Steinbeck is NOT a Nevada writer and this row must not pretend otherwise — his Nevada footprint is a summer job, a divorce and a highway. The pairing is structural, not biographical, and it is offered as a reading. A claim that Carol Steinbeck took a Reno divorce in 1943 circulates and could not be confirmed here; the 1948 decree is the one that checks out.

28. The cleavage is not north against souththe presentch. XIII

    Population shares and the absence of any physical connection between the Truckee and Colorado systems are settled. ⚠ How Washoe sits in the revenue ledger — gaming and sales tax generated overwhelmingly in Clark, with rural counties as net recipients — is argued in good faith by people using different models, and should be carried as a contested question rather than a settled one.

29. The people who actually run the rooms1990s onwardch. XIV

    Local 226’s membership is usually given as about 60,000; Nevada’s Hispanic share is near 30% and the state is majority-minority. ⚠ Both figures move with each estimate — cite the current census release, never this line.

30. Thacker Pass, and the town the company had to buildconstruction from 2023; phase 1 mechanical completion targeted late 2027ch. XIV

    Nevada’s gold share: ~73% of US production in 2023 and ~70% in 2024, per the USGS Mineral Commodity Summaries — the figure moves year to year, so cite the edition. Lithium Americas–GM joint venture 62/38, DOE loan $2.26bn, Bechtel EPCM; phase 1 nameplate 40,000 t/yr; workforce hub 755 modules, ~1,997 rooms, ~850 residents as of March 2026 against ~1,300 on site in mid-May, target 1,800 by year end; ~500 permanent staff over ~40 years. ⚠ Every figure here is a company or programme projection on a project still under construction, and schedules of this kind move — treat the whole row as provisional and re-derive before use. Nevada Copper at Pumpkin Hollow entered bankruptcy in 2024 and its status has moved since; the old Anaconda site at Yerington is a contamination cleanup, not a producer.

31. A labour market with unusual entry conditionsthe camps, and the slope todaych. XIV

    The entry-condition mechanism is structural and well attested across remote extraction economies. ⚠ The relative SIZE of the two populations is asserted here, not measured — I have no figure for what share of any Nevada camp was in flight from something, and neither, I suspect, does anyone.

32. What sits next to a campcurrent, and documented across the western extraction corridorch. XIV

    The association between transient extraction workforce housing and elevated violent crime, with disproportionate harm to Indigenous women, is documented across the western and northern-plains extraction corridors. ⚠ Study designs, geographies and baselines vary and the effect size is genuinely argued — cite a specific study for a specific place, never a general figure, and never attribute it to individuals.

33. Nobody is assigned to the watersnapshot of the NDWR assignment table, August 2026ch. XIV

    Computed from the NDWR Basin Specialist Assignments layer, 256 rows: 83 touching a vacancy, 38 wholly vacant, representing about 12% of mapped basin area. ⚠ THIS IS A SNAPSHOT OF A LIVE ADMINISTRATIVE FILE and vacancies are filled and reopened continuously — the figure dates fast and must be recomputed from the current layer before it is quoted anywhere. It measures assigned staffing, not workload or competence, and a vacant slot does not mean a basin is unadministered.

34. Rural power did not disappear in 1964 — it relocatedReynolds 1964 · TRPA 1969 · AB 413 1979 · Storey ordinance 1999 · 2014 · 2016ch. XIV

    ⚠ PREREGISTERED AND UNTESTED. The dates are real; the causal claim is not established. The falsifier is specified in advance: count county land-use ordinances and federal-land lawsuits by decade before and after 1964. If rural political activity did not migrate, the claim dies. 🔴 AND IT NOW HAS A RIVAL, which is exactly what preregistration is for. The Federal Land Policy and Management Act of 1976 ended the federal disposal policy and made retention permanent — the state that had always assumed it would eventually receive its land was told it never would. The Sagebrush Rebellion follows three years later. That is a shorter lag and a far more direct mechanism than a 1964 apportionment case, and any honest test has to run both. The Reynolds story may still hold for county land-use activity; FLPMA is the better explanation for the federal-land half. Do not merge them — they predict different things.

35. The constraint has flippedthe present, and the first chapter the state cannot write by statutech. XIV

    A forward-looking reading, flagged as one, and the row on this page most likely to be wrong. The strongest objection is real and should be stated with it: NEPA (1970), the Clean Water Act (1972) and the Endangered Species Act (1973) did not exist during the takings above — the cui-ui listing is what gave the tribe a lever in the first place, the 1990 settlement was negotiated rather than seized, and battery recycling genuinely converts a stock toward a flow. Against that: Thacker Pass was litigated by tribes, ranchers and conservation groups and is being built anyway, so participation is not veto; and TRIC is private land inside a county zone designed to engage the federal participatory architecture as little as possible.

36. Zero out of eleventhe strongest result in the volumech. XIV

    Computed by siteLogicResult() over the 28-town register: one-reason 0/11 living, multi-reason 15/17. By reason — resource 3/15, water 10/11, junction 10/11, admin 3/4. ✅ RE-RUN AGAINST AN INDEPENDENT SOURCE and it holds, in a softer form. Matching 27 of the 28 towns to Blevins & Helbock’s US Post Offices dataset — establishment and discontinuance years recorded by the Post Office Department, with no judgement of mine in them — gives a median post-office life of 122 years for multi-reason towns against 63 for one-reason towns, and 15 of 16 still open against 4 of 11. ⚠ SO THE HEADLINE MUST BE QUOTED AS 122-AGAINST-63, OR 15-OF-16 AGAINST 4-OF-11 — NOT as 0 of 11. The zero was measuring my own `status` field. Four one-reason towns still collect mail: Virginia City, Austin, Pioche, Eureka and Goldfield are remnants with post offices. A post office is a lagging indicator — it records whether anybody is still there, not whether the place is a living town. Both measures point the same way and the independent one is more conservative, which is the right direction for a correction to run.

37. Seattle sold shovels. Nevada was the fieldSeattle 1897; Elko and Winnemucca throughout; Fernley 1999ch. XIV

    Seattle's Klondike outfitting role is well documented. Amazon at Fernley from 1999 — the same year as Storey County's preapproval ordinance, which is coincidence and should be reported as one. ⚠ The claim that Elko and Winnemucca survived BECAUSE they outfitted rather than because they are junctions is not separable on this data: both towns are coded water+junction, and outfitting may be a consequence of the junction rather than an independent cause.

38. No port, ever — and the business Nevada has always been in1849 to the present, at four cadencesch. XIV

    Landlocked geography and the ocean-versus-truck freight differential are uncontested. ⚠ The inland-port position is a DERIVATIVE of somebody else's port volumes and moves when they move — Panama capacity, Gulf and East Coast share, near-shoring to Mexico. And it is contested: the Inland Empire, Phoenix, Salt Lake and Las Vegas compete for the same containers. Note too that a large part of the advantage is LEGAL rather than physical — no inventory tax, no corporate income tax, fast permitting — and any state can copy those. Nobody can copy being one day from Los Angeles; the tax half of the moat is shallower than it looks.

39. Specific assets and redeployable onesthe general form of the survival findingch. XIV

    A reframing of the measured result (register: 3/15 against 12/13; post office records: median life 122 years against 63) in the language of asset specificity. ⚠ PREREGISTERED AND UNTESTED IN ITS NEW FORM: the prediction is that single-purpose industrial capital strands at a higher rate than general-purpose infrastructure in the SAME corridor. Thacker Pass states a forty-year life in its own filings; a logistics park at a junction states none. If single-tenant campuses in this corridor survive at the same rate as multi-tenant ones over the next two decades, the reframing is refused and stock-or-flow was the better model after all.

40. Stock or flowthe law the register turns out to be describingch. XIV

    Computed by survivalByResource() over lib/nevada-geo.ts: stock 4 of 8 living, flow 9 of 10. The town register in lib/nevada-towns.ts gives 3 of 15 against 12 of 13 over a larger and differently classified set. ⚠ Both are small-n and both depend on where “remnant” ends and “living” begins — Dayton and Virginia City move the result on their own. The weakest assignment in the table is `trail`, classed as a flow: a trail renews like a river but unlike a river it can be re-routed, and Genoa is a single case propping up that call.

41. Every county line serves a dead economymost formed 1861; Clark 1909; Pershing 1919ch. XIV

    Seventeen counties; the bulk formed 1861; Clark 1909; Pershing 1919 the last; Carson City consolidated as an independent city in 1969 and sits in no county. ⚠ County formation years are the field most likely to be wrong in the geo table — check against the statutes before publishing any of them individually.

42. One habit, not twothe volume's last linech. XIV

    A reading over the whole volume and the row most exposed to being wrong, because it forecasts. Every clause traces to a row above; the forecast in the last sentence does not, and should be read as the claim it is.