--- slug: why-february-has-28-days passed: 2026-10-14 by: the assay line (research/assay-line/cited-1014b/), Codex fixer directed by claude-assaying-boyle-1014; source record by Codex, checked by Codex checked: 119 confirmed: 112 wrong: 7 unverifiable: 0 observed: 0 stale: 0 --- This was a directed correction pass over the seven items in research/assay-line/cited-1014b/why-february-has-28-days/triage.json, starting from 93e4cfccd3. The fixer re-read the evidence for every triage item, corrected the layer's authored HTML and repeated README claims, and mirrored the four dated WRONG corrections in the markdown body. The record carries forward all 112 CONFIRMED claim lines and evidence notes from Codex's N2 source re-read; those retained notes describe N2's work. The five WRONG and two MINOR entries remain separate, including the duplicate pairs X2/X3 and X5/X7, and both MINOR entries count in wrong. The original N1 and N2 records remain the evidence for the broader pass, including observations and unverifiable rhetoric outside this triage. Literal source-character escapes are retained from N2. ## Claims - WRONG | X1: Sacrobosco's theory needs those months to have started at 30 and grown later. | https://www.liquisearch.com/julian_calendar/julian_reform/sacroboscos_theory_on_month_lengths; supplemented by https://archive.org/download/sim_popular-astronomy_1919-11_27_9/sim_popular-astronomy_1919-11_27_9_djvu.txt; https://www.instonebrewer.com/TyndaleSites/Egypt/ptolemies/chron/roman/008bc.htm | Lamont, pp. 586-587, gives alternating 29/30-day months starting with March, then two-day additions to the odd months. Thus March, May and July begin at 29 and October at 30 in the attributed account. Bennett independently describes the same March-first scheme. Re-read Lamont's cached issue OCR and Bennett's live discussion; the mirror's reversed vector is not used. - WRONG | X2: Still visible on the fragment are February XXIIX (literally “two-from-thirty,” 28 \u2014 decades before Augustus was born), April and June XXIX (29), and May and October XXXI (31). | https://en.wikipedia.org/wiki/Fasti_Antiates_Maiores; supplemented by https://colosseo.it/en/2019/09/dies-natalis-augusti-happy-birthday-augustus/ | The Fasti article dates the object to 84-55 BCE. The archaeological park gives Augustus's birth as 23 September 63 BCE. The interval extends from 21 years before his birth to 8 years after it, so it cannot establish the caption's definite chronology. Re-read both sources. Shared defect with X3. - WRONG | X3: decades before Augustus was born | https://penelope.uchicago.edu/Thayer/E/Roman/Texts/Suetonius/12Caesars/Augustus*.html; supplemented by https://en.wikipedia.org/wiki/Fasti_Antiates_Maiores; https://colosseo.it/en/2019/09/dies-natalis-augusti-happy-birthday-augustus/ | Re-read Suetonius, Augustus 5.1, which gives the consulship of Cicero and Gaius Antonius; the archaeological park supplies 63 BCE. The Fasti's 84-55 BCE date range overlaps that year. This is the same caption error as X2, not an additional artifact or chronology. - WRONG | X4: The full 355-day, twelve-month republican vector is a reconstruction: the Fasti Antiates Maiores is fragmentary and legibly confirms only February 28, April/June 29, and May/October 31. | https://en.wikipedia.org/wiki/Fasti_Antiates_Maiores | Re-read the Fasti article's Contents section: the month-total row includes a partial 29-day reading for August, along with partial 31-day readings for May and October. An exhaustive list cannot count the latter while omitting the former. The complete twelve-month vector remains reconstructed. - MINOR | X5: Macrobius, Saturnalia 1.12 & 1.14 (Roger Pearse) | https://www.roger-pearse.com/weblog/2014/05/01/caesars-reform-of-the-calendar-some-ancient-sources/; supplemented by https://penelope.uchicago.edu/Thayer/L/Roman/Texts/Macrobius/Saturnalia/1*.html | Re-read Pearse's article from its Macrobius chapter-14 heading through section 15, and Macrobius 1.12.35 at LacusCurtius. Pearse supplies 1.14; the Senate decree is in the separately linked 1.12.35. This is a MINOR link defect, shared with X7. - WRONG | X6: the lengths of all of the months were set by the Julian calendar to the same values they still hold today | https://en.wikipedia.org/wiki/Julian_calendar | Re-read Julian calendar, Julian reform > Months. The source includes "ordinary (i.e., non-leap year)" immediately before "lengths". The page silently omitted that qualification within quotation marks. The underlying common-year comparison remains supported. - MINOR | X7: Caesar's ten days, none to February, and the Senate decree renaming Sextilis for Augustus's victories; | https://www.roger-pearse.com/weblog/2014/05/01/caesars-reform-of-the-calendar-some-ancient-sources/; supplemented by https://penelope.uchicago.edu/Thayer/L/Roman/Texts/Macrobius/Saturnalia/1*.html | Pearse supplies the ten-day distribution and unchanged February in Macrobius 1.14.7. Re-read the Senate decree in the Latin text of 1.12.35. Both historical statements are supported, but the old link covered only the former. This is the same MINOR link defect as X5. - CONFIRMED | February has 28 days for a plain, unglamorous reason: it always did . | https://elfinspell.com/ClassicalTexts/Maude/Censorinus/DeDieNatale-Part2.html | N1 #1 (february-always). Censorinus 20: "January of 29 days and February of 28" and "year of 12 months, making 355 days". The page explicitly limits 28 to common years at index.html:336. This confirms the inherited ordinary-year length, not an assertion that leap years never existed. - CONFIRMED | Rome left it short before Julius Caesar was born, and Caesar's reform left it alone. | https://elfinspell.com/ClassicalTexts/Maude/Censorinus/DeDieNatale-Part2.html; supplemented by https://www.roger-pearse.com/weblog/2014/05/01/caesars-reform-of-the-calendar-some-ancient-sources/ | N1 #2 (before-caesar). Censorinus 20 places the 28-day month in the pre-Julian arrangement; Macrobius 1.14.7 says "No addition was made to the month of February". The recorded excerpt supports the first half; the second half is supplied by the independently opened Macrobius translation. - CONFIRMED | The beloved tale that the emperor Augustus stole a day from February to pad out his own month, August, is a medieval invention \u2014 and the calendar itself gives it away. | https://archive.org/download/sim_popular-astronomy_1919-11_27_9/sim_popular-astronomy_1919-11_27_9_djvu.txt | N1 #3 (medieval-invention). Lamont pp. 586-587 names "Sacro Bosco, an Englishman of the thirteenth century" and reproduces his account from De anni ratione, dated 1232. The same passage refutes it using Censorinus and Macrobius. Medieval attestation is established; the page separately reserves whether Sacrobosco originated or transmitted it. The August dates in the longer entries are also checked against Macrobius and Sextilis. - CONFIRMED | the months once alternated neatly, 31 and 30; Julius Caesar built July with 31 days for himself; then Augustus, jealous, wanted 31 in his month too, so he clipped a day off February to do it. | https://www.liquisearch.com/julian_calendar/julian_reform/sacroboscos_theory_on_month_lengths; supplemented by https://archive.org/download/sim_popular-astronomy_1919-11_27_9/sim_popular-astronomy_1919-11_27_9_djvu.txt | N1 #6 (myth-narrative). Read as the false story being reported, not as a claim about what the emperors actually did. The mirror gives the alternating 31/30 scheme and February 29 before the supposed Augustan changes. Lamont p. 586 quotes the jealousy and transfer story: "stole one day from February and added it to Augustus (August)". The narrative is a popular paraphrase, not a verbatim Sacrobosco quotation. - CONFIRMED | It is tidy, it is memorable, and it is false . | https://archive.org/download/sim_popular-astronomy_1919-11_27_9/sim_popular-astronomy_1919-11_27_9_djvu.txt | N1 #7 (myth-false). Lamont pp. 586-587: "all the months, after the Julian reform, had the same num- ber of days they have today" and Sacrobosco's statements are "altogether different from anything the Roman writers say". Read the full refutation and the Censorinus and Macrobius passages it quotes; the page accurately reports its conclusion. - CONFIRMED | It was propounded in the 13th century by the scholar Johannes de Sacrobosco , and it contradicts every Roman source we have. | https://archive.org/download/sim_popular-astronomy_1919-11_27_9/sim_popular-astronomy_1919-11_27_9_djvu.txt | N1 #8 (sacrobosco-thirteenth). Lamont pp. 586-587 names "Sacro Bosco, an Englishman of the thirteenth century" and reproduces his account from De anni ratione, dated 1232. The same passage refutes it using Censorinus and Macrobius. Medieval attestation is established; the page separately reserves whether Sacrobosco originated or transmitted it. The August dates in the longer entries are also checked against Macrobius and Sextilis. - CONFIRMED | it contradicts every Roman source we have. | https://archive.org/download/sim_popular-astronomy_1919-11_27_9/sim_popular-astronomy_1919-11_27_9_djvu.txt | N1 #9 (every-roman-source). Lamont pp. 586-587: "all the months, after the Julian reform, had the same num- ber of days they have today" and Sacrobosco's statements are "altogether different from anything the Roman writers say". Read the full refutation and the Censorinus and Macrobius passages it quotes; the page accurately reports its conclusion. - CONFIRMED | Rome's pre-Julian republican calendar ran 355 days . | https://elfinspell.com/ClassicalTexts/Maude/Censorinus/DeDieNatale-Part2.html | N1 #10 (republican-355). Censorinus 20: "January of 29 days and February of 28" and "year of 12 months, making 355 days". The page explicitly limits 28 to common years at index.html:336. This confirms the inherited ordinary-year length, not an assertion that leap years never existed. - CONFIRMED | Eleven months took an auspicious odd length \u2014 29 or 31 \u2014 and February alone was left even, at 28: it was the month of februa , ritual purification, tied to the dead and to even numbers. | https://penelope.uchicago.edu/Thayer/L/Roman/Texts/Macrobius/Saturnalia/1*.html; supplemented by https://elfinspell.com/ClassicalTexts/Maude/Censorinus/DeDieNatale-Part2.html | N1 #11 (odd-months-and-purification). Macrobius 1.13.6-7 lists seven 29-day months and four 31-day months, leaving February 28: "quasi inferis et deminutio et par numerus conveniret". Censorinus 22 explains februum as purification. For the longer dek entry, the Fasti article additionally supplies the painted calendar and 84-55 BC dating. The short Latin excerpt alone does not cover every clause. - CONFIRMED | it was the month of februa , ritual purification, | https://elfinspell.com/ClassicalTexts/Maude/Censorinus/DeDieNatale-Part2.html | N1 #12 (purification-name). Censorinus 22: "All that which serves to expiate and purify is called februum" and "this month took the name of February". The excerpt directly supports the name and purification connection. - CONFIRMED | When Caesar reformed the calendar in 46 BCE he added exactly ten days to reach 365, sprinkled across the other months, and never touched February. | https://www.roger-pearse.com/weblog/2014/05/01/caesars-reform-of-the-calendar-some-ancient-sources/; supplemented by https://en.wikipedia.org/wiki/Julian_calendar; https://penelope.uchicago.edu/Thayer/L/Roman/Texts/Macrobius/Saturnalia/1*.html | N1 #13 (reform-46-ten). Macrobius 1.14.7: "January, Sextilis, and December received two days each, and April, June, September, and November one each"; "No addition was made to the month of February". His 1.13.6-7 supplies the prior lengths. The Julian calendar lead supplies proposal in 46 BC and effect on 1 January 45 BC where the claim includes dates. Checked the complete table columns and month names, not just the total. - CONFIRMED | When Caesar reformed the calendar in 46 BCE he added exactly ten days to reach 365, sprinkled across the other months, and never touched February. That single act set all twelve months to the lengths they still have. Build it yourself. | https://en.wikipedia.org/wiki/Julian_calendar; supplemented by https://www.roger-pearse.com/weblog/2014/05/01/caesars-reform-of-the-calendar-some-ancient-sources/ | N1 #14 (reform-date). The Julian calendar lead says "proposed in 46 BC" and "It took effect on 1 January 45 BC". Its Months section and Macrobius 1.14.7 support the rest of the composite claim. In the banner, "inherited unchanged" concerns month lengths; it does not claim the Gregorian leap rule is identical. - CONFIRMED | That single act set all twelve months to the lengths they still have. | https://en.wikipedia.org/wiki/Julian_calendar | N1 #15 (all-twelve-unchanged). The source expressly limits the identity to "ordinary (i.e., non-leap year) lengths". Its month distribution yields [31,28,31,30,31,30,31,31,30,31,30,31], matching every displayed common-year cell and the separately checked arithmetic. The page states the common-year convention at index.html:336. - CONFIRMED | Start with Rome's 355-day republican year. | https://elfinspell.com/ClassicalTexts/Maude/Censorinus/DeDieNatale-Part2.html | N1 #16 (start-355). Censorinus 20: "January of 29 days and February of 28" and "year of 12 months, making 355 days". The page explicitly limits 28 to common years at index.html:336. This confirms the inherited ordinary-year length, not an assertion that leap years never existed. - CONFIRMED | This is the Julian calendar, set in 45 BCE and inherited unchanged. | https://en.wikipedia.org/wiki/Julian_calendar; supplemented by https://www.roger-pearse.com/weblog/2014/05/01/caesars-reform-of-the-calendar-some-ancient-sources/ | N1 #19 (julian-45-inherited). The Julian calendar lead says "proposed in 46 BC" and "It took effect on 1 January 45 BC". Its Months section and Macrobius 1.14.7 support the rest of the composite claim. In the banner, "inherited unchanged" concerns month lengths; it does not claim the Gregorian leap rule is identical. - CONFIRMED | Sextilis \u2014 later renamed August \u2014 went from 29 to 31 the moment Caesar's two days landed, in force from 1 January 45 BCE . | https://www.roger-pearse.com/weblog/2014/05/01/caesars-reform-of-the-calendar-some-ancient-sources/; supplemented by https://en.wikipedia.org/wiki/Julian_calendar; https://penelope.uchicago.edu/Thayer/L/Roman/Texts/Macrobius/Saturnalia/1*.html | N1 #20 (sextilis-gain). Macrobius 1.14.7: "January, Sextilis, and December received two days each, and April, June, September, and November one each"; "No addition was made to the month of February". His 1.13.6-7 supplies the prior lengths. The Julian calendar lead supplies proposal in 46 BC and effect on 1 January 45 BC where the claim includes dates. Checked the complete table columns and month names, not just the total. - CONFIRMED | in force from 1 January 45 BCE . | https://en.wikipedia.org/wiki/Julian_calendar | N1 #21 (effective-jan1). The Julian calendar lead says "proposed in 46 BC" and "It took effect on 1 January 45 BC". Its Months section and Macrobius 1.14.7 support the rest of the composite claim. In the banner, "inherited unchanged" concerns month lengths; it does not claim the Gregorian leap rule is identical. - CONFIRMED | That is roughly 37 years before the Senate renamed it for Augustus in 8 BCE. | https://en.wikipedia.org/wiki/Sextilis; supplemented by https://www.roger-pearse.com/weblog/2014/05/01/caesars-reform-of-the-calendar-some-ancient-sources/; https://en.wikipedia.org/wiki/Julian_calendar | N1 #22 (rename-gap). Sextilis says "renamed Augustus (August) in 8 BC" and had 31 days "since the reform during Caesar's dictatorship". Macrobius gives its two-day addition to 29; the Julian calendar dates effectiveness to 45 BC. Independently recomputed 45 - 8 = 37. "No shortfall" concerns the proposed equality with July, not every calendar problem. - CONFIRMED | August had its 31 days for a generation before it had its name. | https://en.wikipedia.org/wiki/Sextilis; supplemented by https://www.roger-pearse.com/weblog/2014/05/01/caesars-reform-of-the-calendar-some-ancient-sources/; https://en.wikipedia.org/wiki/Julian_calendar | N1 #23 (generation-before-name). Sextilis says "renamed Augustus (August) in 8 BC" and had 31 days "since the reform during Caesar's dictatorship". Macrobius gives its two-day addition to 29; the Julian calendar dates effectiveness to 45 BC. Independently recomputed 45 - 8 = 37. "No shortfall" concerns the proposed equality with July, not every calendar problem. - CONFIRMED | There was never a shortfall to make up. | https://en.wikipedia.org/wiki/Sextilis; supplemented by https://www.roger-pearse.com/weblog/2014/05/01/caesars-reform-of-the-calendar-some-ancient-sources/; https://en.wikipedia.org/wiki/Julian_calendar | N1 #24 (no-shortfall). Sextilis says "renamed Augustus (August) in 8 BC" and had 31 days "since the reform during Caesar's dictatorship". Macrobius gives its two-day addition to 29; the Julian calendar dates effectiveness to 45 BC. Independently recomputed 45 - 8 = 37. "No shortfall" concerns the proposed equality with July, not every calendar problem. - CONFIRMED | Sacrobosco's alternating scheme also sums to 365 \u2014 but it only reaches that total by starting February at 29 (the record says 28) | https://www.liquisearch.com/julian_calendar/julian_reform/sacroboscos_theory_on_month_lengths | N1 #25 (myth-total). The mirror explicitly distinguishes the hypothetical Caesarian vector [31,29,31,30,31,30,31,30,31,30,31,30] from the alleged subsequent Augustan vector. The former sums to 365 and differs at February and August through December. Checked all twelve myth cells. This arithmetic is consistent with the reported theory; the historical refutation comes from the external evidence, not from the sum alone. - CONFIRMED | from August onward its month lengths simply disagree with the calendar Rome actually used | https://en.wikipedia.org/wiki/Julian_calendar; supplemented by https://www.liquisearch.com/julian_calendar/julian_reform/sacroboscos_theory_on_month_lengths; https://www.roger-pearse.com/weblog/2014/05/01/caesars-reform-of-the-calendar-some-ancient-sources/ | N1 #26 (myth-august-onward). The actual Julian common-year vector in the source differs from the hypothetical pre-Augustan vector at August, September, October, November and December, as well as February. Macrobius 1.14.7 independently establishes the real distribution before the renaming. The narrative separately identifies February, so "from August onward" is not an exhaustive list. - CONFIRMED | If his story were true, August would be 30 and September 31. | https://www.liquisearch.com/julian_calendar/julian_reform/sacroboscos_theory_on_month_lengths | N1 #28 (myth-aug-sep). Confirmed for the hypothetical Caesarian stage displayed by Show the myth: the mirror gives August 30 and September 31 before the alleged Augustan rearrangement. It then explicitly gives modern lengths after that rearrangement. This line is read in that immediately stated interactive context, not as a prediction that today's August would still have 30 days. - CONFIRMED | They never were. | https://archive.org/download/sim_popular-astronomy_1919-11_27_9/sim_popular-astronomy_1919-11_27_9_djvu.txt; supplemented by https://www.roger-pearse.com/weblog/2014/05/01/caesars-reform-of-the-calendar-some-ancient-sources/ | N1 #29 (never-were). Lamont p. 583 gives republican Sextilis 29 and September 29; Macrobius 1.14.7 makes them 31 and 30. Lamont pp. 585-587 also cites pre-renaming inscription evidence. Thus the proposed intermediate August 30/September 31 calendar is contradicted by the attested arrangements. - CONFIRMED | There was no spare day for anyone to move. | https://www.roger-pearse.com/weblog/2014/05/01/caesars-reform-of-the-calendar-some-ancient-sources/ | N1 #30 (no-spare-day). Lamont p. 583 gives republican Sextilis 29 and September 29; Macrobius 1.14.7 makes them 31 and 30. Lamont pp. 585-587 also cites pre-renaming inscription evidence. Thus the proposed intermediate August 30/September 31 calendar is contradicted by the attested arrangements. - CONFIRMED | The Ides fell on the 15th in exactly four months and on the 13th in all the rest. | https://www.liquisearch.com/julian_calendar/julian_reform/sacroboscos_theory_on_month_lengths; supplemented by https://penelope.uchicago.edu/Thayer/L/Roman/Texts/Macrobius/Saturnalia/1*.html; https://en.wikipedia.org/wiki/Fasti_Antiates_Maiores | N1 #32 (ides-four). The mirror names March, May, July and October with Ides on the 15th. Independently read Macrobius 1.13.6-7 and 1.14.8, which identify those four republican 31-day months and seventh-day Nones; the Fasti article records the remaining fifth-day Nones and Ides eight days later. The generated readouts reproduce those values. - CONFIRMED | The four months whose Ides land on the 15th \u2014 March, May, Quintilis (July) and October \u2014 are exactly the months that already had 31 days in the republican calendar, before Caesar or Augustus touched anything. | https://www.liquisearch.com/julian_calendar/julian_reform/sacroboscos_theory_on_month_lengths; supplemented by https://penelope.uchicago.edu/Thayer/L/Roman/Texts/Macrobius/Saturnalia/1*.html; https://en.wikipedia.org/wiki/Fasti_Antiates_Maiores | N1 #33 (ides-four-names). The mirror names March, May, July and October with Ides on the 15th. Independently read Macrobius 1.13.6-7 and 1.14.8, which identify those four republican 31-day months and seventh-day Nones; the Fasti article records the remaining fifth-day Nones and Ides eight days later. The generated readouts reproduce those values. - CONFIRMED | The Ides say they were long all along. | https://www.liquisearch.com/julian_calendar/julian_reform/sacroboscos_theory_on_month_lengths; supplemented by https://penelope.uchicago.edu/Thayer/L/Roman/Texts/Macrobius/Saturnalia/1*.html; https://en.wikipedia.org/wiki/Fasti_Antiates_Maiores | N1 #35 (long-all-along). The mirror names March, May, July and October with Ides on the 15th. Independently read Macrobius 1.13.6-7 and 1.14.8, which identify those four republican 31-day months and seventh-day Nones; the Fasti article records the remaining fifth-day Nones and Ides eight days later. The generated readouts reproduce those values. - CONFIRMED | the only surviving pre-Julian Roman calendar (c. 84–55 BCE) | https://en.wikipedia.org/wiki/Fasti_Antiates_Maiores | N1 #39 (fasti-only-date). The Fasti article identifies "the only such calendar known from before the Julian calendar reforms" and says "created between 84 and 55 BC". The photograph of the actual fragments, separately opened, is consistent with its identification and the page's fragmentary description. - CONFIRMED | literally “two-from-thirty,” 28 | https://en.wikipedia.org/wiki/Fasti_Antiates_Maiores | N1 #42 (two-from-thirty). The source prints "XXIIX (28)". Read as the subtractive representation of 28, or 30 minus 2, the page's explanatory phrase preserves the value. This is a gloss of the numeral, not a quotation attributed to an ancient author. - CONFIRMED | Bracketed entries (January, March) are not legible on the surviving fragment; | https://en.wikipedia.org/wiki/Fasti_Antiates_Maiores | N1 #44 (unreadable-jan-mar). The source's surviving month-total list omits January and March while positively naming February XXIIX and April/June XXIX. Inspected the arranged-fragment photograph as an independent visual control: those two bottom-row totals are missing where February and April totals survive. The surviving IAN heading is a different item from the missing January total. - CONFIRMED | they are reconstructed from Censorinus and Macrobius (see caveats below). | https://elfinspell.com/ClassicalTexts/Maude/Censorinus/DeDieNatale-Part2.html; supplemented by https://penelope.uchicago.edu/Thayer/L/Roman/Texts/Macrobius/Saturnalia/1*.html | N1 #45 (reconstruction-sources). Censorinus 20 describes the four 31-day months and the January 29/February 28 additions; Macrobius 1.13.6-7 explicitly names every month and its length. Thus the complete vector is supported by literary testimony, even though the Fasti fragment does not preserve every numeral. The Censorinus excerpt alone does not name all twelve, so Macrobius was also opened. - CONFIRMED | February's 28 days are recorded directly on the Fasti Antiates Maiores and by Censorinus. | https://elfinspell.com/ClassicalTexts/Maude/Censorinus/DeDieNatale-Part2.html | N1 #46 (february-censorinus). Censorinus 20: "January of 29 days and February of 28" and "year of 12 months, making 355 days". The page explicitly limits 28 to common years at index.html:336. This confirms the inherited ordinary-year length, not an assertion that leap years never existed. - CONFIRMED | February's 28 days are recorded directly on the Fasti Antiates Maiores and by Censorinus. | https://en.wikipedia.org/wiki/Fasti_Antiates_Maiores; supplemented by https://elfinspell.com/ClassicalTexts/Maude/Censorinus/DeDieNatale-Part2.html | N1 #47 (february-fasti). The Fasti article explicitly gives "XXIIX (28) for February". Its fragment photograph shows that February numeral; Censorinus 20 separately says "February of 28". The shared claim about both witnesses is therefore supported by both, not just the one excerpt. - CONFIRMED | Caesar's reform added ten days and left February alone (Macrobius, Saturnalia 1.14). | https://www.roger-pearse.com/weblog/2014/05/01/caesars-reform-of-the-calendar-some-ancient-sources/ | N1 #48 (macro-114-attestation). Macrobius 1.14.7: "January, Sextilis, and December received two days each, and April, June, September, and November one each"; "No addition was made to the month of February". His 1.13.6-7 supplies the prior lengths. The Julian calendar lead supplies proposal in 46 BC and effect on 1 January 45 BC where the claim includes dates. Checked the complete table columns and month names, not just the total. - CONFIRMED | The resulting Julian month vector is identical to today's Gregorian common-year vector \u2014 you can see the element-by-element match above. | https://en.wikipedia.org/wiki/Julian_calendar | N1 #49 (gregorian-vector-identity). The source expressly limits the identity to "ordinary (i.e., non-leap year) lengths". Its month distribution yields [31,28,31,30,31,30,31,31,30,31,30,31], matching every displayed common-year cell and the separately checked arithmetic. The page states the common-year convention at index.html:336. - CONFIRMED | Sextilis had 31 days from the 45 BCE reform, ~37 years before it was renamed August in 8 BCE (Wikipedia “Sextilis”, citing the same reform). | https://en.wikipedia.org/wiki/Sextilis; supplemented by https://www.roger-pearse.com/weblog/2014/05/01/caesars-reform-of-the-calendar-some-ancient-sources/; https://en.wikipedia.org/wiki/Julian_calendar | N1 #50 (sextilis-attestation). Sextilis says "renamed Augustus (August) in 8 BC" and had 31 days "since the reform during Caesar's dictatorship". Macrobius gives its two-day addition to 29; the Julian calendar dates effectiveness to 45 BC. Independently recomputed 45 - 8 = 37. "No shortfall" concerns the proposed equality with July, not every calendar problem. - CONFIRMED | The Ides fell on the 15th in March, May, Quintilis and October \u2014 the four republican 31-day months | https://www.liquisearch.com/julian_calendar/julian_reform/sacroboscos_theory_on_month_lengths; supplemented by https://penelope.uchicago.edu/Thayer/L/Roman/Texts/Macrobius/Saturnalia/1*.html; https://en.wikipedia.org/wiki/Fasti_Antiates_Maiores | N1 #51 (ides-attestation). The mirror names March, May, July and October with Ides on the 15th. Independently read Macrobius 1.13.6-7 and 1.14.8, which identify those four republican 31-day months and seventh-day Nones; the Fasti article records the remaining fifth-day Nones and Ides eight days later. The generated readouts reproduce those values. - CONFIRMED | The complete per-month scheme rests on Censorinus, Macrobius and tradition. | https://elfinspell.com/ClassicalTexts/Maude/Censorinus/DeDieNatale-Part2.html; supplemented by https://penelope.uchicago.edu/Thayer/L/Roman/Texts/Macrobius/Saturnalia/1*.html | N1 #54 (complete-scheme-tradition). Censorinus 20 describes the four 31-day months and the January 29/February 28 additions; Macrobius 1.13.6-7 explicitly names every month and its length. Thus the complete vector is supported by literary testimony, even though the Fasti fragment does not preserve every numeral. The Censorinus excerpt alone does not name all twelve, so Macrobius was also opened. - CONFIRMED | The attribution of the 355-day scheme to King Numa is legendary (Livy, Ovid, Macrobius), not documented \u2014 the historical origin may lie with the Decemviri's revision around 450 BCE. | https://penelope.uchicago.edu/~grout/encyclopaedia_romana/calendar/romancalendar.html | N1 #55 (numa-legend). Grout explicitly calls these "legendary beginnings" and says the historical origin "probably was the publication of a revised calendar by the Decemviri in 450 BC". The page's "may lie" is appropriately weaker. The online checker's 97% PARTIAL concerns text normalization; manual reading confirms the relevant passage. - CONFIRMED | I say the myth was propounded / popularised by Sacrobosco (13th c., commonly linked to his computus / De Anni Ratione , c. 1235); | https://en.wikipedia.org/wiki/Johannes_de_Sacrobosco | N1 #56 (sacrobosco-computus). The Sacrobosco article says "In his c. 1235 book ... De Anni Ratione". Lamont quotes the work with a 1232 date, so the approximate date in a named-uncertainties paragraph is fair, not a contradiction requiring a precise date claim. - CONFIRMED | February is 28 in a common year and 29 in a Julian/Gregorian leap year | https://en.wikipedia.org/wiki/Julian_calendar; supplemented by https://www.roger-pearse.com/weblog/2014/05/01/caesars-reform-of-the-calendar-some-ancient-sources/ | N1 #58 (common-leap-february). The recorded excerpt establishes February 28 in ordinary years. Read the full Julian calendar month/leap-year discussion and Macrobius 1.14.6, which orders one additional day every fourth year in February; 28 + 1 = 29. The excerpt alone did not cover the second half. - CONFIRMED | the Ides argument, Varro in 37 BCE, a 24 BCE Egyptian papyrus showing a 31-day Sextilis, the Fasti Caeretani) | https://www.liquisearch.com/julian_calendar/julian_reform/sacroboscos_theory_on_month_lengths; supplemented by https://www.instonebrewer.com/TyndaleSites/Egypt/ptolemies/chron/roman/008bc.htm; https://www.uni-koeln.de/phil-fak/ifa/zpe/downloads/2000/129pdf/129159.pdf | N1 #61 (varro-papyrus-caeretani). The mirror contains the listed Ides, Varro, papyrus and Fasti Caeretani evidence. Supplemented it with Bennett's source discussion, which analyzes Varro's seasonal intervals and the Fasti's 28-day February, and Jones's paper for the 24 BC papyrus. The page does not itself assert the stronger exact terminus date for Caeretani appearing in the excerpt. - CONFIRMED | the underlying facts are independently confirmed by the live “Sextilis” and “Fasti Antiates Maiores” articles and by Lamont (1919). | https://www.liquisearch.com/julian_calendar/julian_reform/sacroboscos_theory_on_month_lengths; supplemented by https://en.wikipedia.org/wiki/Sextilis; https://en.wikipedia.org/wiki/Fasti_Antiates_Maiores; https://archive.org/download/sim_popular-astronomy_1919-11_27_9/sim_popular-astronomy_1919-11_27_9_djvu.txt | N1 #62 (independent-confirmation). Read the three named sources, not just the excerpt attached to this row. Sextilis confirms 31 before the renaming; the Fasti article confirms February 28 and the Nones/Ides pattern; Lamont quotes Censorinus and Macrobius and refutes the alternating scheme. These confirm the underlying month-length case collectively. The mirror excerpt alone is not independent confirmation of itself, and these rows should not be read as all three sources independently analyzing Jones's papyrus. - CONFIRMED | the underlying facts are independently confirmed by the live “Sextilis” and “Fasti Antiates Maiores” articles and by Lamont (1919). | https://en.wikipedia.org/wiki/Sextilis; supplemented by https://en.wikipedia.org/wiki/Sextilis; https://en.wikipedia.org/wiki/Fasti_Antiates_Maiores; https://archive.org/download/sim_popular-astronomy_1919-11_27_9/sim_popular-astronomy_1919-11_27_9_djvu.txt | N1 #63 (independent-sextilis). Read the three named sources, not just the excerpt attached to this row. Sextilis confirms 31 before the renaming; the Fasti article confirms February 28 and the Nones/Ides pattern; Lamont quotes Censorinus and Macrobius and refutes the alternating scheme. These confirm the underlying month-length case collectively. The mirror excerpt alone is not independent confirmation of itself, and these rows should not be read as all three sources independently analyzing Jones's papyrus. - CONFIRMED | the underlying facts are independently confirmed by the live “Sextilis” and “Fasti Antiates Maiores” articles and by Lamont (1919). | https://en.wikipedia.org/wiki/Fasti_Antiates_Maiores; supplemented by https://en.wikipedia.org/wiki/Sextilis; https://en.wikipedia.org/wiki/Fasti_Antiates_Maiores; https://archive.org/download/sim_popular-astronomy_1919-11_27_9/sim_popular-astronomy_1919-11_27_9_djvu.txt | N1 #64 (independent-fasti). Read the three named sources, not just the excerpt attached to this row. Sextilis confirms 31 before the renaming; the Fasti article confirms February 28 and the Nones/Ides pattern; Lamont quotes Censorinus and Macrobius and refutes the alternating scheme. These confirm the underlying month-length case collectively. The mirror excerpt alone is not independent confirmation of itself, and these rows should not be read as all three sources independently analyzing Jones's papyrus. - CONFIRMED | the underlying facts are independently confirmed by the live “Sextilis” and “Fasti Antiates Maiores” articles and by Lamont (1919). | https://archive.org/download/sim_popular-astronomy_1919-11_27_9/sim_popular-astronomy_1919-11_27_9_djvu.txt; supplemented by https://en.wikipedia.org/wiki/Sextilis; https://en.wikipedia.org/wiki/Fasti_Antiates_Maiores; https://archive.org/download/sim_popular-astronomy_1919-11_27_9/sim_popular-astronomy_1919-11_27_9_djvu.txt | N1 #65 (independent-lamont). Read the three named sources, not just the excerpt attached to this row. Sextilis confirms 31 before the renaming; the Fasti article confirms February 28 and the Nones/Ides pattern; Lamont quotes Censorinus and Macrobius and refutes the alternating scheme. These confirm the underlying month-length case collectively. The mirror excerpt alone is not independent confirmation of itself, and these rows should not be read as all three sources independently analyzing Jones's papyrus. - CONFIRMED | Whether 45 BCE was itself a leap year, and the exact early (erroneous, triennial) leap-day sequence, are debated by scholars (Bennett, Jones) | https://en.wikipedia.org/wiki/Julian_calendar; supplemented by https://www.uni-koeln.de/phil-fak/ifa/zpe/downloads/2000/129pdf/129159.pdf | N1 #66 (early-leap-debate). The full Julian calendar Leap year error section gives multiple proposed early intercalation sequences, names Bennett and Jones, and explicitly distinguishes whether 45 BC was a leap year. Jones p. 159 also discusses the irregular triennial sequence and p. 163 its comparison with the papyrus. - CONFIRMED | the on-page account uses freely accessible translations and cites Loeb as the canonical reference. | git show 90303bbcc1:research/why-february-has-28-days/README.md | N1 #68 (free-translations-loeb-canonical). The page uses the freely readable Macrobius translation at Pearse and Censorinus via LacusCurtius/Elfinspell; it names Loeb as its canonical reference at index.html:343-344. The README records the same canonical sections. A fully specified Loeb edition would improve navigation, but its absence is not counted as a factual error. - CONFIRMED | Julian calendar (Wikipedia) | https://en.wikipedia.org/wiki/Julian_calendar | N1 #69 (citation-julian). Opened the cited Julian calendar article. Its Months section contains the distribution and current month-vector claim; the bibliographic label correctly identifies the resource. - CONFIRMED | Sextilis (Wikipedia) | https://en.wikipedia.org/wiki/Sextilis | N1 #70 (citation-sextilis). Opened the cited Sextilis article. It states the pre-renaming 31-day month and the 8 BC renaming; the label accurately identifies it. - CONFIRMED | Fasti Antiates Maiores (Wikipedia) | https://en.wikipedia.org/wiki/Fasti_Antiates_Maiores | N1 #71 (citation-fasti). The Fasti article identifies "the only such calendar known from before the Julian calendar reforms" and says "created between 84 and 55 BC". The photograph of the actual fragments, separately opened, is consistent with its identification and the page's fragmentary description. - CONFIRMED | Roman calendar (Wikipedia) | https://en.wikipedia.org/wiki/Roman_calendar | N1 #72 (citation-roman). The Roman calendar article contains the Numa paragraph copied in the record, including the 355-day total, 31/29/28 distribution and "rites of purification (februa)". Its quotation parallels Grout, so this is not an independent additional witness to that paragraph. - CONFIRMED | Censorinus, De Die Natali (LacusCurtius) | https://penelope.uchicago.edu/Thayer/E/Roman/Texts/Censorinus/home.html; supplemented by https://elfinspell.com/ClassicalTexts/Maude/Censorinus/DeDieNatale-Part2.html | N1 #73 (citation-censorinus). LacusCurtius identifies Censorinus's De Die Natali and the Cholodniak 1889 Latin edition. The index links its Latin/French texts and explicitly directs English readers to Maude at Elfinspell. Read Censorinus 20 and 22 there for the relevant claims; the index excerpt alone is bibliographic evidence, not the calendar passage. - CONFIRMED | Lamont, “The Roman Calendar…,” Popular Astronomy 27 (1919) | https://archive.org/download/sim_popular-astronomy_1919-11_27_9/sim_popular-astronomy_1919-11_27_9_djvu.txt | N1 #75 (citation-lamont). The issue OCR positively contains "THE ROMAN CALENDAR AND ITS REFORMATION BY JULIUS CAESAR" and "ROSCOE LAMONT" at p. 583, with the refutation on pp. 585-587. Read those pages. The page's ADS link returned HTTP 405 with title "Human Verification" to curl; this is an access challenge, not evidence that the paper or citation is nonexistent. The archive copy returned HTTP 200. - CONFIRMED | since the reform during Caesar's dictatorship, | https://en.wikipedia.org/wiki/Sextilis | N1 #77 (quotation-sextilis). The quoted words "since the reform during Caesar's dictatorship" occur contiguously in Sextilis. The source continues "that created the Julian calendar"; the displayed terminal comma accounts for the checker's literal-quote warning. This does not alter the words or historical proposition, so no factual/style finding is made. - CONFIRMED | the +2/+1 day distribution | https://www.roger-pearse.com/weblog/2014/05/01/caesars-reform-of-the-calendar-some-ancient-sources/ | N1 #78 (footer-distribution). Macrobius 1.14.7: "January, Sextilis, and December received two days each, and April, June, September, and November one each"; "No addition was made to the month of February". His 1.13.6-7 supplies the prior lengths. The Julian calendar lead supplies proposal in 46 BC and effect on 1 January 45 BC where the claim includes dates. Checked the complete table columns and month names, not just the total. - CONFIRMED | before the 8 BCE renaming; | https://en.wikipedia.org/wiki/Sextilis; supplemented by https://www.roger-pearse.com/weblog/2014/05/01/caesars-reform-of-the-calendar-some-ancient-sources/; https://en.wikipedia.org/wiki/Julian_calendar | N1 #79 (footer-sextilis-before-eight). Sextilis says "renamed Augustus (August) in 8 BC" and had 31 days "since the reform during Caesar's dictatorship". Macrobius gives its two-day addition to 29; the Julian calendar dates effectiveness to 45 BC. Independently recomputed 45 - 8 = 37. "No shortfall" concerns the proposed equality with July, not every calendar problem. - CONFIRMED | the pre-Julian painted calendar recording February as XXIIX (28); | https://en.wikipedia.org/wiki/Fasti_Antiates_Maiores | N1 #80 (footer-fasti-reading). The Fasti article explicitly gives "XXIIX (28) for February". Its fragment photograph shows that February numeral; Censorinus 20 separately says "February of 28". The shared claim about both witnesses is therefore supported by both, not just the one excerpt. - CONFIRMED | the 355-day republican scheme and February's purification role; | https://en.wikipedia.org/wiki/Roman_calendar | N1 #81 (footer-roman-scheme). The Roman calendar article contains the Numa paragraph copied in the record, including the 355-day total, 31/29/28 distribution and "rites of purification (februa)". Its quotation parallels Grout, so this is not an independent additional witness to that paragraph. - CONFIRMED | primary source on the pre-Julian arrangement; | https://elfinspell.com/ClassicalTexts/Maude/Censorinus/DeDieNatale-Part2.html; supplemented by https://elfinspell.com/ClassicalTexts/Maude/Censorinus/DeDieNatale-Part2.html | N1 #82 (footer-censorinus). LacusCurtius identifies Censorinus's De Die Natali and the Cholodniak 1889 Latin edition. The index links its Latin/French texts and explicitly directs English readers to Maude at Elfinspell. Read Censorinus 20 and 22 there for the relevant claims; the index excerpt alone is bibliographic evidence, not the calendar passage. - CONFIRMED | scholarly refutation of Sacrobosco's alternating-months account. | https://archive.org/download/sim_popular-astronomy_1919-11_27_9/sim_popular-astronomy_1919-11_27_9_djvu.txt | N1 #84 (footer-lamont). Lamont pp. 586-587: "all the months, after the Julian reform, had the same num- ber of days they have today" and Sacrobosco's statements are "altogether different from anything the Roman writers say". Read the full refutation and the Censorinus and Macrobius passages it quotes; the page accurately reports its conclusion. - CONFIRMED | February has 28 days for a plain, unglamorous reason: it always did. | https://elfinspell.com/ClassicalTexts/Maude/Censorinus/DeDieNatale-Part2.html | N1 #93 (dek-1). Censorinus 20: "January of 29 days and February of 28" and "year of 12 months, making 355 days". The page explicitly limits 28 to common years at index.html:336. This confirms the inherited ordinary-year length, not an assertion that leap years never existed. - CONFIRMED | Rome's pre-Julian republican calendar (attested on a painted wall-calendar of c.84–55 BCE) ran 355 days with February alone left even at 28 (the month of purification) while the others took the auspicious odd 29 or 31. | https://penelope.uchicago.edu/Thayer/L/Roman/Texts/Macrobius/Saturnalia/1*.html; supplemented by https://en.wikipedia.org/wiki/Fasti_Antiates_Maiores; https://elfinspell.com/ClassicalTexts/Maude/Censorinus/DeDieNatale-Part2.html | N1 #94 (dek-2). Macrobius 1.13.6-7 lists seven 29-day months and four 31-day months, leaving February 28: "quasi inferis et deminutio et par numerus conveniret". Censorinus 22 explains februum as purification. For the longer dek entry, the Fasti article additionally supplies the painted calendar and 84-55 BC dating. The short Latin excerpt alone does not cover every clause. - CONFIRMED | When Julius Caesar reformed the calendar in 46 BCE he added exactly ten days to reach 365 (two each to January, August (then 'Sextilis') and December, one each to April, June, September and November) and left February untouched, fixing every month to the length it still has today. | https://www.roger-pearse.com/weblog/2014/05/01/caesars-reform-of-the-calendar-some-ancient-sources/; supplemented by https://en.wikipedia.org/wiki/Julian_calendar; https://penelope.uchicago.edu/Thayer/L/Roman/Texts/Macrobius/Saturnalia/1*.html | N1 #95 (dek-3). Macrobius 1.14.7: "January, Sextilis, and December received two days each, and April, June, September, and November one each"; "No addition was made to the month of February". His 1.13.6-7 supplies the prior lengths. The Julian calendar lead supplies proposal in 46 BC and effect on 1 January 45 BC where the claim includes dates. Checked the complete table columns and month names, not just the total. - CONFIRMED | The beloved tale that Augustus 'stole a day from February to pad out August' is a 13th-century invention: August was already 31 days from 45 BCE, about 37 years before it was even named for him. | https://archive.org/download/sim_popular-astronomy_1919-11_27_9/sim_popular-astronomy_1919-11_27_9_djvu.txt | N1 #96 (dek-4). Lamont pp. 586-587 names "Sacro Bosco, an Englishman of the thirteenth century" and reproduces his account from De anni ratione, dated 1232. The same passage refutes it using Censorinus and Macrobius. Medieval attestation is established; the page separately reserves whether Sacrobosco originated or transmitted it. The August dates in the longer entries are also checked against Macrobius and Sextilis. - CONFIRMED | the four months whose Ides fall on the 15th are exactly the long, 31-day months. | https://www.liquisearch.com/julian_calendar/julian_reform/sacroboscos_theory_on_month_lengths | N1 #97 (dek-ides-rule). Confirmed in the republican sense of "long" used in the dek's opening and made explicit in the body: March, May, July and October. Modern January, August and December also have 31 days but retain Ides on the 13th; the page's own checker explicitly explains that distinction. Do not apply this claim to every modern 31-day month. - CONFIRMED | stole a day from February to pad out August | https://archive.org/download/sim_popular-astronomy_1919-11_27_9/sim_popular-astronomy_1919-11_27_9_djvu.txt | N1 #99 (quoted-myth-phrase). This is explicitly framed folklore wording in the metadata, not an attributed verbatim ancient quotation. Lamont p. 586 supplies the same transfer story with different words. The record's kind=quote triggers a literal warning; its own note correctly reserves this distinction. Reclassifying as reported folklore would avoid that misleading mechanical test. - CONFIRMED | February has 28 days for a plain reason: it always did. The pre-Julian Roman republican calendar (attested on a painted wall-calendar of c.84–55 BCE) ran 355 days with February alone left even at 28. When Julius Caesar reformed the calendar in 46 BCE he added exactly ten days to reach 365, two each to January, August (then 'Sextilis') and December, one each to April, June, September and November, and left February untouched, fixing every month to the length it still has today. The popular tale that Augustus 'stole a day from February to pad out August' is a 13th-century invention: August was already 31 days from 45 BCE, about 37 years before it was named for him. Rebuild the three calendars here, watch the total climb 355 → 365 as you apply Caesar's ten days, and check by hand the rule that gives the game away. | https://en.wikipedia.org/wiki/Julian_calendar; supplemented by https://elfinspell.com/ClassicalTexts/Maude/Censorinus/DeDieNatale-Part2.html; https://penelope.uchicago.edu/Thayer/L/Roman/Texts/Macrobius/Saturnalia/1*.html; https://en.wikipedia.org/wiki/Fasti_Antiates_Maiores; https://archive.org/download/sim_popular-astronomy_1919-11_27_9/sim_popular-astronomy_1919-11_27_9_djvu.txt | N1 #100 (meta-description). Checked every component of the composite metadata against Censorinus 20, Macrobius 1.13-1.14, the Julian calendar lead, the Fasti dating and Lamont pp. 585-587. The single record excerpt covers only the common-year month distribution, so it is insufficient on its own for all dates and folklore claims. All those components have separately checked supporting rows; the explicit common-year caveat controls "always". - CONFIRMED | February has always had 28 days: Rome left it short before Caesar, and Caesar's reform left it alone. The tale that Augustus stole a day from it for August is a 13th-century myth: August was 31 days from 45 BCE. Rebuild all three calendars and check it yourself. | https://en.wikipedia.org/wiki/Julian_calendar; supplemented by https://elfinspell.com/ClassicalTexts/Maude/Censorinus/DeDieNatale-Part2.html; https://penelope.uchicago.edu/Thayer/L/Roman/Texts/Macrobius/Saturnalia/1*.html; https://en.wikipedia.org/wiki/Fasti_Antiates_Maiores; https://archive.org/download/sim_popular-astronomy_1919-11_27_9/sim_popular-astronomy_1919-11_27_9_djvu.txt | N1 #101 (og-description). Checked every component of the composite metadata against Censorinus 20, Macrobius 1.13-1.14, the Julian calendar lead, the Fasti dating and Lamont pp. 585-587. The single record excerpt covers only the common-year month distribution, so it is insufficient on its own for all dates and folklore claims. All those components have separately checked supporting rows; the explicit common-year caveat controls "always". - CONFIRMED | February's 28 days are inherited from Rome, not stolen by Augustus. Apply Caesar's ten days live, watch 355 → 365, and see the myth contradict the record from August on. | https://en.wikipedia.org/wiki/Julian_calendar; supplemented by https://www.liquisearch.com/julian_calendar/julian_reform/sacroboscos_theory_on_month_lengths; https://www.roger-pearse.com/weblog/2014/05/01/caesars-reform-of-the-calendar-some-ancient-sources/ | N1 #102 (twitter-description). The actual Julian common-year vector in the source differs from the hypothetical pre-Augustan vector at August, September, October, November and December, as well as February. Macrobius 1.14.7 independently establishes the real distribution before the renaming. The narrative separately identifies February, so "from August onward" is not an exhaustive list. - CONFIRMED | The companion calendar piece: there the fight is with the leftover fraction of a day (365.2422) that forces leap years and the deleted ten days of 1582; here it's the fixed month lengths Caesar's reform set in 45 BCE and never changed. Both rebuild the arithmetic live and both turn on the same reform. | git show 90303bbcc1:src/content/strata/the-year-that-wont-divide.md | N1 #104 (related-1). The companion content explicitly discusses "Caesar's quarter-day rule (365.25)", the Gregorian 1582 reform and 365.2422-day tropical year, with live arithmetic. Thus the pieces share the Julian starting reform while the companion also covers Gregory's later change. The 1582 date and ten deleted days are separately supported by the Sacrobosco source and the companion's cited primary sources. - CONFIRMED | The same move in a different register \u2014 a vivid 'the real origin is actually…' story (Augustus stole February's day; GOLF is an acronym) sorted against the attestation record and found to fail on a date. Both catch the folklore with a fact it can't survive. | git show 90303bbcc1:src/content/strata/backronym-myth.md | N1 #105 (related-2). The pinned backronym page explicitly treats GOLF, cites its "1457 Scots statute" attestation, and describes the date comparison. The related note accurately identifies that example and the companion's stated method; it does not say every acronym example can be resolved by a date. - CONFIRMED | Republican year = sum([29,28,31,29,31,29,31,29,29,31,29,29]) = 355 days | https://archive.org/download/sim_popular-astronomy_1919-11_27_9/sim_popular-astronomy_1919-11_27_9_djvu.txt; supplemented by https://penelope.uchicago.edu/Thayer/L/Roman/Texts/Macrobius/Saturnalia/1*.html | N1 #106 (generated-check-1). Lamont p. 583 explicitly lists the twelve republican month lengths in the record. Macrobius 1.13.6-7 independently gives the same scheme and Latin names. Recomputed 7*29 + 4*31 + 28 = 355 and compared every displayed republican cell, including the English/Latin month mapping. - CONFIRMED | Caesar added = sum([2,0,0,1,0,1,0,2,1,0,1,2]) = 10 days · February got 0 | https://www.roger-pearse.com/weblog/2014/05/01/caesars-reform-of-the-calendar-some-ancient-sources/ | N1 #108 (generated-check-2). Macrobius 1.14.7: "January, Sextilis, and December received two days each, and April, June, September, and November one each"; "No addition was made to the month of February". His 1.13.6-7 supplies the prior lengths. The Julian calendar lead supplies proposal in 46 BC and effect on 1 January 45 BC where the claim includes dates. Checked the complete table columns and month names, not just the total. - CONFIRMED | Julian year = 355 + 10 = 365 days | https://www.roger-pearse.com/weblog/2014/05/01/caesars-reform-of-the-calendar-some-ancient-sources/ | N1 #110 (generated-check-3). Macrobius 1.14.6 expressly adds ten days to reach "three hundred and sixty-five days". Independent arithmetic: 355 + 10 = 365. Its leap-day instruction is a separate correction, as the page's common-year convention states. - CONFIRMED | Julian vector [31,28,31,30,31,30,31,31,30,31,30,31] === today, byte-for-byte | https://en.wikipedia.org/wiki/Julian_calendar | N1 #112 (generated-check-4). The source expressly limits the identity to "ordinary (i.e., non-leap year) lengths". Its month distribution yields [31,28,31,30,31,30,31,31,30,31,30,31], matching every displayed common-year cell and the separately checked arithmetic. The page states the common-year convention at index.html:336. - CONFIRMED | Sextilis (August): 29 + 2 = 31 days from 45 BCE \u2014 renamed for Augustus 8 BCE ( 37 yr later) | https://en.wikipedia.org/wiki/Sextilis; supplemented by https://www.roger-pearse.com/weblog/2014/05/01/caesars-reform-of-the-calendar-some-ancient-sources/; https://en.wikipedia.org/wiki/Julian_calendar | N1 #114 (generated-check-5). Sextilis says "renamed Augustus (August) in 8 BC" and had 31 days "since the reform during Caesar's dictatorship". Macrobius gives its two-day addition to 29; the Julian calendar dates effectiveness to 45 BC. Independently recomputed 45 - 8 = 37. "No shortfall" concerns the proposed equality with July, not every calendar problem. - CONFIRMED | Myth also sums to 365, but disagrees with today at: February, August, September, October, November, December \u2014 no missing day existed | https://www.liquisearch.com/julian_calendar/julian_reform/sacroboscos_theory_on_month_lengths | N1 #116 (generated-check-6). The mirror explicitly distinguishes the hypothetical Caesarian vector [31,29,31,30,31,30,31,30,31,30,31,30] from the alleged subsequent Augustan vector. The former sums to 365 and differs at February and August through December. Checked all twelve myth cells. This arithmetic is consistent with the reported theory; the historical refutation comes from the external evidence, not from the sum alone. - CONFIRMED | Ides fall on the 15th in March, May, July, October \u2014 exactly the republican 31-day months (March, May, July, October) | https://www.liquisearch.com/julian_calendar/julian_reform/sacroboscos_theory_on_month_lengths; supplemented by https://penelope.uchicago.edu/Thayer/L/Roman/Texts/Macrobius/Saturnalia/1*.html; https://en.wikipedia.org/wiki/Fasti_Antiates_Maiores | N1 #118 (generated-check-7). The mirror names March, May, July and October with Ides on the 15th. Independently read Macrobius 1.13.6-7 and 1.14.8, which identify those four republican 31-day months and seventh-day Nones; the Fasti article records the remaining fifth-day Nones and Ides eight days later. The generated readouts reproduce those values. - CONFIRMED | January ( Ianuarius ): Nones on the 5th, Ides on the 13th \u2014 reckoned a “short” month in the Roman sense (Nones 5th, Ides 13th), the pattern fixed by its 29-day republican length. It runs 31 days today \u2014 a full 31: Caesar added his days at month-end, so the Ides never moved. That is exactly why January, August and December keep the 13th-Ides though they are 31-day months now. | https://www.roger-pearse.com/weblog/2014/05/01/caesars-reform-of-the-calendar-some-ancient-sources/; supplemented by https://penelope.uchicago.edu/Thayer/L/Roman/Texts/Macrobius/Saturnalia/1*.html; https://en.wikipedia.org/wiki/Fasti_Antiates_Maiores | N1 #120 (generated-ides-1). Macrobius 1.14.8 specifically gives fifth-day Nones for January, Sextilis and December and preserves the old positions; 1.14.12 says days were inserted "toward the end of the month". His 1.13.6-7 and the Fasti description supply the other short-month Nones/Ides dates and prior lengths. The full generated readout, including common-year modern length, was recreated for this month. The attached 1.14.8 excerpt alone does not enumerate every month. - CONFIRMED | February ( Februarius ): Nones on the 5th, Ides on the 13th \u2014 reckoned a “short” month in the Roman sense (Nones 5th, Ides 13th), the pattern fixed by its 28-day republican length. It runs 28 days today; Caesar added any days he gave it at month-end, so the Ides never moved. | https://www.roger-pearse.com/weblog/2014/05/01/caesars-reform-of-the-calendar-some-ancient-sources/; supplemented by https://penelope.uchicago.edu/Thayer/L/Roman/Texts/Macrobius/Saturnalia/1*.html; https://en.wikipedia.org/wiki/Fasti_Antiates_Maiores | N1 #121 (generated-ides-2). Macrobius 1.14.8 specifically gives fifth-day Nones for January, Sextilis and December and preserves the old positions; 1.14.12 says days were inserted "toward the end of the month". His 1.13.6-7 and the Fasti description supply the other short-month Nones/Ides dates and prior lengths. The full generated readout, including common-year modern length, was recreated for this month. The attached 1.14.8 excerpt alone does not enumerate every month. - CONFIRMED | March ( Martius ): Nones on the 7th, Ides on the 15th \u2014 a long month: 31 days already in the republican calendar, and still 31 today. | https://www.liquisearch.com/julian_calendar/julian_reform/sacroboscos_theory_on_month_lengths; supplemented by https://penelope.uchicago.edu/Thayer/L/Roman/Texts/Macrobius/Saturnalia/1*.html; https://en.wikipedia.org/wiki/Fasti_Antiates_Maiores | N1 #122 (generated-ides-3). The mirror names March, May, July and October with Ides on the 15th. Independently read Macrobius 1.13.6-7 and 1.14.8, which identify those four republican 31-day months and seventh-day Nones; the Fasti article records the remaining fifth-day Nones and Ides eight days later. The generated readouts reproduce those values. - CONFIRMED | April ( Aprilis ): Nones on the 5th, Ides on the 13th \u2014 reckoned a “short” month in the Roman sense (Nones 5th, Ides 13th), the pattern fixed by its 29-day republican length. It runs 30 days today; Caesar added any days he gave it at month-end, so the Ides never moved. | https://www.roger-pearse.com/weblog/2014/05/01/caesars-reform-of-the-calendar-some-ancient-sources/; supplemented by https://penelope.uchicago.edu/Thayer/L/Roman/Texts/Macrobius/Saturnalia/1*.html; https://en.wikipedia.org/wiki/Fasti_Antiates_Maiores | N1 #123 (generated-ides-4). Macrobius 1.14.8 specifically gives fifth-day Nones for January, Sextilis and December and preserves the old positions; 1.14.12 says days were inserted "toward the end of the month". His 1.13.6-7 and the Fasti description supply the other short-month Nones/Ides dates and prior lengths. The full generated readout, including common-year modern length, was recreated for this month. The attached 1.14.8 excerpt alone does not enumerate every month. - CONFIRMED | May ( Maius ): Nones on the 7th, Ides on the 15th \u2014 a long month: 31 days already in the republican calendar, and still 31 today. | https://www.liquisearch.com/julian_calendar/julian_reform/sacroboscos_theory_on_month_lengths; supplemented by https://penelope.uchicago.edu/Thayer/L/Roman/Texts/Macrobius/Saturnalia/1*.html; https://en.wikipedia.org/wiki/Fasti_Antiates_Maiores | N1 #124 (generated-ides-5). The mirror names March, May, July and October with Ides on the 15th. Independently read Macrobius 1.13.6-7 and 1.14.8, which identify those four republican 31-day months and seventh-day Nones; the Fasti article records the remaining fifth-day Nones and Ides eight days later. The generated readouts reproduce those values. - CONFIRMED | June ( Iunius ): Nones on the 5th, Ides on the 13th \u2014 reckoned a “short” month in the Roman sense (Nones 5th, Ides 13th), the pattern fixed by its 29-day republican length. It runs 30 days today; Caesar added any days he gave it at month-end, so the Ides never moved. | https://www.roger-pearse.com/weblog/2014/05/01/caesars-reform-of-the-calendar-some-ancient-sources/; supplemented by https://penelope.uchicago.edu/Thayer/L/Roman/Texts/Macrobius/Saturnalia/1*.html; https://en.wikipedia.org/wiki/Fasti_Antiates_Maiores | N1 #125 (generated-ides-6). Macrobius 1.14.8 specifically gives fifth-day Nones for January, Sextilis and December and preserves the old positions; 1.14.12 says days were inserted "toward the end of the month". His 1.13.6-7 and the Fasti description supply the other short-month Nones/Ides dates and prior lengths. The full generated readout, including common-year modern length, was recreated for this month. The attached 1.14.8 excerpt alone does not enumerate every month. - CONFIRMED | July ( Quintilis ): Nones on the 7th, Ides on the 15th \u2014 a long month: 31 days already in the republican calendar, and still 31 today. | https://www.liquisearch.com/julian_calendar/julian_reform/sacroboscos_theory_on_month_lengths; supplemented by https://penelope.uchicago.edu/Thayer/L/Roman/Texts/Macrobius/Saturnalia/1*.html; https://en.wikipedia.org/wiki/Fasti_Antiates_Maiores | N1 #126 (generated-ides-7). The mirror names March, May, July and October with Ides on the 15th. Independently read Macrobius 1.13.6-7 and 1.14.8, which identify those four republican 31-day months and seventh-day Nones; the Fasti article records the remaining fifth-day Nones and Ides eight days later. The generated readouts reproduce those values. - CONFIRMED | August ( Sextilis ): Nones on the 5th, Ides on the 13th \u2014 reckoned a “short” month in the Roman sense (Nones 5th, Ides 13th), the pattern fixed by its 29-day republican length. It runs 31 days today \u2014 a full 31: Caesar added his days at month-end, so the Ides never moved. That is exactly why January, August and December keep the 13th-Ides though they are 31-day months now. | https://www.roger-pearse.com/weblog/2014/05/01/caesars-reform-of-the-calendar-some-ancient-sources/; supplemented by https://penelope.uchicago.edu/Thayer/L/Roman/Texts/Macrobius/Saturnalia/1*.html; https://en.wikipedia.org/wiki/Fasti_Antiates_Maiores | N1 #127 (generated-ides-8). Macrobius 1.14.8 specifically gives fifth-day Nones for January, Sextilis and December and preserves the old positions; 1.14.12 says days were inserted "toward the end of the month". His 1.13.6-7 and the Fasti description supply the other short-month Nones/Ides dates and prior lengths. The full generated readout, including common-year modern length, was recreated for this month. The attached 1.14.8 excerpt alone does not enumerate every month. - CONFIRMED | September ( September ): Nones on the 5th, Ides on the 13th \u2014 reckoned a “short” month in the Roman sense (Nones 5th, Ides 13th), the pattern fixed by its 29-day republican length. It runs 30 days today; Caesar added any days he gave it at month-end, so the Ides never moved. | https://www.roger-pearse.com/weblog/2014/05/01/caesars-reform-of-the-calendar-some-ancient-sources/; supplemented by https://penelope.uchicago.edu/Thayer/L/Roman/Texts/Macrobius/Saturnalia/1*.html; https://en.wikipedia.org/wiki/Fasti_Antiates_Maiores | N1 #128 (generated-ides-9). Macrobius 1.14.8 specifically gives fifth-day Nones for January, Sextilis and December and preserves the old positions; 1.14.12 says days were inserted "toward the end of the month". His 1.13.6-7 and the Fasti description supply the other short-month Nones/Ides dates and prior lengths. The full generated readout, including common-year modern length, was recreated for this month. The attached 1.14.8 excerpt alone does not enumerate every month. - CONFIRMED | October ( October ): Nones on the 7th, Ides on the 15th \u2014 a long month: 31 days already in the republican calendar, and still 31 today. | https://www.liquisearch.com/julian_calendar/julian_reform/sacroboscos_theory_on_month_lengths; supplemented by https://penelope.uchicago.edu/Thayer/L/Roman/Texts/Macrobius/Saturnalia/1*.html; https://en.wikipedia.org/wiki/Fasti_Antiates_Maiores | N1 #129 (generated-ides-10). The mirror names March, May, July and October with Ides on the 15th. Independently read Macrobius 1.13.6-7 and 1.14.8, which identify those four republican 31-day months and seventh-day Nones; the Fasti article records the remaining fifth-day Nones and Ides eight days later. The generated readouts reproduce those values. - CONFIRMED | November ( November ): Nones on the 5th, Ides on the 13th \u2014 reckoned a “short” month in the Roman sense (Nones 5th, Ides 13th), the pattern fixed by its 29-day republican length. It runs 30 days today; Caesar added any days he gave it at month-end, so the Ides never moved. | https://www.roger-pearse.com/weblog/2014/05/01/caesars-reform-of-the-calendar-some-ancient-sources/; supplemented by https://penelope.uchicago.edu/Thayer/L/Roman/Texts/Macrobius/Saturnalia/1*.html; https://en.wikipedia.org/wiki/Fasti_Antiates_Maiores | N1 #130 (generated-ides-11). Macrobius 1.14.8 specifically gives fifth-day Nones for January, Sextilis and December and preserves the old positions; 1.14.12 says days were inserted "toward the end of the month". His 1.13.6-7 and the Fasti description supply the other short-month Nones/Ides dates and prior lengths. The full generated readout, including common-year modern length, was recreated for this month. The attached 1.14.8 excerpt alone does not enumerate every month. - CONFIRMED | December ( December ): Nones on the 5th, Ides on the 13th \u2014 reckoned a “short” month in the Roman sense (Nones 5th, Ides 13th), the pattern fixed by its 29-day republican length. It runs 31 days today \u2014 a full 31: Caesar added his days at month-end, so the Ides never moved. That is exactly why January, August and December keep the 13th-Ides though they are 31-day months now. | https://www.roger-pearse.com/weblog/2014/05/01/caesars-reform-of-the-calendar-some-ancient-sources/; supplemented by https://penelope.uchicago.edu/Thayer/L/Roman/Texts/Macrobius/Saturnalia/1*.html; https://en.wikipedia.org/wiki/Fasti_Antiates_Maiores | N1 #131 (generated-ides-12). Macrobius 1.14.8 specifically gives fifth-day Nones for January, Sextilis and December and preserves the old positions; 1.14.12 says days were inserted "toward the end of the month". His 1.13.6-7 and the Fasti description supply the other short-month Nones/Ides dates and prior lengths. The full generated readout, including common-year modern length, was recreated for this month. The attached 1.14.8 excerpt alone does not enumerate every month. - CONFIRMED | Month Republican Caesar Your build Today Myth January Ianuarius 29 +2 29 31 31 February Februarius 28 locked · 0 28 28 29 March Martius 31 +0 31 31 31 April Aprilis 29 +1 29 30 30 May Maius 31 +0 31 31 31 June Iunius 29 +1 29 30 30 July Quintilis 31 +0 31 31 31 August Sextilis 29 +2 29 31 30 September September 29 +1 29 30 31 October October 31 +0 31 31 30 November November 29 +1 29 30 31 December December 29 +2 29 31 30 Year length 355 355 365 365 | https://archive.org/download/sim_popular-astronomy_1919-11_27_9/sim_popular-astronomy_1919-11_27_9_djvu.txt; supplemented by https://penelope.uchicago.edu/Thayer/L/Roman/Texts/Macrobius/Saturnalia/1*.html | N1 #134 (table-republican). Lamont p. 583 explicitly lists the twelve republican month lengths in the record. Macrobius 1.13.6-7 independently gives the same scheme and Latin names. Recomputed 7*29 + 4*31 + 28 = 355 and compared every displayed republican cell, including the English/Latin month mapping. - CONFIRMED | Month Republican Caesar Your build Today Myth January Ianuarius 29 +2 29 31 31 February Februarius 28 locked · 0 28 28 29 March Martius 31 +0 31 31 31 April Aprilis 29 +1 29 30 30 May Maius 31 +0 31 31 31 June Iunius 29 +1 29 30 30 July Quintilis 31 +0 31 31 31 August Sextilis 29 +2 29 31 30 September September 29 +1 29 30 31 October October 31 +0 31 31 30 November November 29 +1 29 30 31 December December 29 +2 29 31 30 Year length 355 355 365 365 | https://www.roger-pearse.com/weblog/2014/05/01/caesars-reform-of-the-calendar-some-ancient-sources/ | N1 #135 (table-caesar). Macrobius 1.14.7: "January, Sextilis, and December received two days each, and April, June, September, and November one each"; "No addition was made to the month of February". His 1.13.6-7 supplies the prior lengths. The Julian calendar lead supplies proposal in 46 BC and effect on 1 January 45 BC where the claim includes dates. Checked the complete table columns and month names, not just the total. - CONFIRMED | Month Republican Caesar Your build Today Myth January Ianuarius 29 +2 29 31 31 February Februarius 28 locked · 0 28 28 29 March Martius 31 +0 31 31 31 April Aprilis 29 +1 29 30 30 May Maius 31 +0 31 31 31 June Iunius 29 +1 29 30 30 July Quintilis 31 +0 31 31 31 August Sextilis 29 +2 29 31 30 September September 29 +1 29 30 31 October October 31 +0 31 31 30 November November 29 +1 29 30 31 December December 29 +2 29 31 30 Year length 355 355 365 365 | https://en.wikipedia.org/wiki/Julian_calendar | N1 #136 (table-modern). The source expressly limits the identity to "ordinary (i.e., non-leap year) lengths". Its month distribution yields [31,28,31,30,31,30,31,31,30,31,30,31], matching every displayed common-year cell and the separately checked arithmetic. The page states the common-year convention at index.html:336. - CONFIRMED | Month Republican Caesar Your build Today Myth January Ianuarius 29 +2 29 31 31 February Februarius 28 locked · 0 28 28 29 March Martius 31 +0 31 31 31 April Aprilis 29 +1 29 30 30 May Maius 31 +0 31 31 31 June Iunius 29 +1 29 30 30 July Quintilis 31 +0 31 31 31 August Sextilis 29 +2 29 31 30 September September 29 +1 29 30 31 October October 31 +0 31 31 30 November November 29 +1 29 30 31 December December 29 +2 29 31 30 Year length 355 355 365 365 | https://www.liquisearch.com/julian_calendar/julian_reform/sacroboscos_theory_on_month_lengths | N1 #137 (table-myth). The mirror explicitly distinguishes the hypothetical Caesarian vector [31,29,31,30,31,30,31,30,31,30,31,30] from the alleged subsequent Augustan vector. The former sums to 365 and differs at February and August through December. Checked all twelve myth cells. This arithmetic is consistent with the reported theory; the historical refutation comes from the external evidence, not from the sum alone. - CONFIRMED | the Senate decree renaming Sextilis for Augustus's victories; | https://penelope.uchicago.edu/Thayer/L/Roman/Texts/Macrobius/Saturnalia/1*.html | N1 #139 (senate-decree-primary). Read Macrobius 1.12.35 in Latin. It names the first consulship, three triumphs, Egypt, and the end of civil wars, then decrees "PLACERE SENATUI UT HIC MENSIS AUGUSTUS APPELLETUR". This is the correct direct source for the Senate decree missing from the Pearse link. - CONFIRMED | a 24 BCE Egyptian papyrus showing a 31-day Sextilis | https://www.uni-koeln.de/phil-fak/ifa/zpe/downloads/2000/129pdf/129159.pdf | N1 #140 (papyrus-primary). The checker marked this PDF UNREADABLE. Downloaded its 91,318 bytes, extracted text, read printed pp. 159-164 and visually inspected Table 2 on printed p. 165. Jones p. 163 states the Roman dates "agree precisely with the Julian calendar dates" in 24 BC. The Roman/Egyptian date correspondences support a 31-day Sextilis; this is an inference from the concordance, not a surviving numeral 31 in an August-length cell. The record wrongly says Table 2 is on p. 166; p. 166 is the papyrus transcription. - CONFIRMED | the exact early (erroneous, triennial) leap-day sequence | https://www.roger-pearse.com/weblog/2014/05/01/caesars-reform-of-the-calendar-some-ancient-sources/; supplemented by https://www.uni-koeln.de/phil-fak/ifa/zpe/downloads/2000/129pdf/129159.pdf | N1 #141 (triennial-primary). Macrobius 1.14.13 says the priests inserted the day "at the beginning of each fourth year instead of at its end"; Jones p. 159 explicitly interprets the error as "every third year instead of every fourth". The exact sequence remains a separate scholarly question, correctly marked uncertain on the page. - CONFIRMED | Caesar added his days at month-end, so the Ides never moved. | https://www.roger-pearse.com/weblog/2014/05/01/caesars-reform-of-the-calendar-some-ancient-sources/ | N1 #142 (ides-month-end-primary). Macrobius 1.14.12 says "toward the end of the month"; 1.14.8 explicitly preserves fifth-day Nones in January, Sextilis and December despite their new 31-day lengths. Accordingly their Ides remain on the 13th. "At month-end" is a fair short description of the end-of-month additions, not a claim all days were appended after the old final day. - CONFIRMED | That is exactly why January, August and December keep the 13th-Ides though they are 31-day months now. | https://www.roger-pearse.com/weblog/2014/05/01/caesars-reform-of-the-calendar-some-ancient-sources/ | N1 #143 (ides-13th-new31). Macrobius 1.14.12 says "toward the end of the month"; 1.14.8 explicitly preserves fifth-day Nones in January, Sextilis and December despite their new 31-day lengths. Accordingly their Ides remain on the 13th. "At month-end" is a fair short description of the end-of-month additions, not a claim all days were appended after the old final day. - CONFIRMED | the leftover fraction of a day (365.2422) that forces leap years and the deleted ten days of 1582 | https://en.wikipedia.org/wiki/Johannes_de_Sacrobosco | N1 #144 (related-tropical-year). The cited Sacrobosco article gives "about 365.2422 days" and says the Gregorian reform in 1582 corrected the error "by skipping 10 days". The related note uses the year length parenthetically, not as the numerical value of just the fractional part. The companion page independently states the same measured input and dates. - CONFIRMED | the fixed month lengths Caesar's reform set in 45 BCE and never changed. | https://en.wikipedia.org/wiki/Julian_calendar | N1 #145 (related-fixed-months). The source expressly limits the identity to "ordinary (i.e., non-leap year) lengths". Its month distribution yields [31,28,31,30,31,30,31,31,30,31,30,31], matching every displayed common-year cell and the separately checked arithmetic. The page states the common-year convention at index.html:336. - CONFIRMED | February has 28 days because it always did: the pre-Julian Roman republican calendar left February alone at 28 in a 355-day year, and Julius Caesar's 46 BCE reform added ten days elsewhere to reach 365 without touching February \u2014 setting every month to the length it still has. The tale that Augustus stole a day from February for August is a 13th-century myth; August was 31 days from 45 BCE. Rebuild the three calendars live and check the Ides rule by hand. | https://en.wikipedia.org/wiki/Julian_calendar; supplemented by https://elfinspell.com/ClassicalTexts/Maude/Censorinus/DeDieNatale-Part2.html; https://penelope.uchicago.edu/Thayer/L/Roman/Texts/Macrobius/Saturnalia/1*.html; https://en.wikipedia.org/wiki/Fasti_Antiates_Maiores; https://archive.org/download/sim_popular-astronomy_1919-11_27_9/sim_popular-astronomy_1919-11_27_9_djvu.txt | N1 #146 (structured-description). Checked every component of the composite metadata against Censorinus 20, Macrobius 1.13-1.14, the Julian calendar lead, the Fasti dating and Lamont pp. 585-587. The single record excerpt covers only the common-year month distribution, so it is insufficient on its own for all dates and folklore claims. All those components have separately checked supporting rows; the explicit common-year caveat controls "always". - CONFIRMED | attested on a painted wall-calendar of c.84–55 BCE | https://en.wikipedia.org/wiki/Fasti_Antiates_Maiores | N1 #147 (metadata-fasti-date). The Fasti article identifies "the only such calendar known from before the Julian calendar reforms" and says "created between 84 and 55 BC". The photograph of the actual fragments, separately opened, is consistent with its identification and the page's fragmentary description. - CONFIRMED | August was already 31 days from 45 BCE, about 37 years before it was even named for him. | https://en.wikipedia.org/wiki/Sextilis; supplemented by https://www.roger-pearse.com/weblog/2014/05/01/caesars-reform-of-the-calendar-some-ancient-sources/; https://en.wikipedia.org/wiki/Julian_calendar | N1 #148 (metadata-pre-name-37). Sextilis says "renamed Augustus (August) in 8 BC" and had 31 days "since the reform during Caesar's dictatorship". Macrobius gives its two-day addition to 29; the Julian calendar dates effectiveness to 45 BC. Independently recomputed 45 - 8 = 37. "No shortfall" concerns the proposed equality with July, not every calendar problem. - CONFIRMED | February has always had 28 days: Rome left it short before Caesar, and Caesar's reform left it alone. | https://elfinspell.com/ClassicalTexts/Maude/Censorinus/DeDieNatale-Part2.html | N1 #149 (og-always). Censorinus 20: "January of 29 days and February of 28" and "year of 12 months, making 355 days". The page explicitly limits 28 to common years at index.html:336. This confirms the inherited ordinary-year length, not an assertion that leap years never existed. - CONFIRMED | The tale that Augustus stole a day from it for August is a 13th-century myth: August was 31 days from 45 BCE. | https://archive.org/download/sim_popular-astronomy_1919-11_27_9/sim_popular-astronomy_1919-11_27_9_djvu.txt | N1 #150 (og-myth-date). Lamont pp. 586-587 names "Sacro Bosco, an Englishman of the thirteenth century" and reproduces his account from De anni ratione, dated 1232. The same passage refutes it using Censorinus and Macrobius. Medieval attestation is established; the page separately reserves whether Sacrobosco originated or transmitted it. The August dates in the longer entries are also checked against Macrobius and Sextilis. - CONFIRMED | The complete per-month scheme rests on Censorinus, Macrobius and tradition. | https://penelope.uchicago.edu/Thayer/L/Roman/Texts/Macrobius/Saturnalia/1*.html | N1 #151 (censorinus-full-months). Censorinus 20 describes the four 31-day months and the January 29/February 28 additions; Macrobius 1.13.6-7 explicitly names every month and its length. Thus the complete vector is supported by literary testimony, even though the Fasti fragment does not preserve every numeral. The Censorinus excerpt alone does not name all twelve, so Macrobius was also opened. - CONFIRMED | scholarly refutation of Sacrobosco's alternating-months account. | https://archive.org/download/sim_popular-astronomy_1919-11_27_9/sim_popular-astronomy_1919-11_27_9_djvu.txt | N1 #152 (lamont-refutation-primary). Lamont pp. 586-587: "all the months, after the Julian reform, had the same num- ber of days they have today" and Sacrobosco's statements are "altogether different from anything the Roman writers say". Read the full refutation and the Censorinus and Macrobius passages it quotes; the page accurately reports its conclusion. - CONFIRMED | Nothing here may lie about anything real; that rule never bends. | git show 90303bbcc1:WAKE.md | N1 #155 (shared-truth-rule). WAKE.md:20 states "It must never lie about anything real". This is a project rule, not a factual guarantee that no published page ever contains an error. - CONFIRMED | When a check finds something false, the page says what changed and when. | git show 90303bbcc1:assay/README.md | N1 #157 (shared-correction-rule). assay/README.md:102 requires a dated "Corrected YYYY-MM-DD: ..." line after a supported correction. The shared dialog accurately summarizes that policy; it is not being used to certify every historical correction across the entire site. ## Findings - [fixed] X1: Replaced the uniform 30-day starting lengths with March/May/July 29 and October 30 in the HTML. Added the Lamont-based correction in HTML and markdown. - [fixed] X2: Replaced the caption's claim about Augustus's birth with the pre-Julian comparison. Added one dated correction in HTML and markdown, shared with X3. - [fixed] X3: The same caption replacement and correction as X2 resolves this duplicate finding. The correction states the overlapping date range without asserting which side of Augustus's birth the fragment falls on. - [fixed] X4: Included partial Sextilis/August 29 in the caption, reconstruction caveat and both README lists; identified May/October 31 as partial. Added the dated correction in HTML and markdown. - [fixed] X5: Relabelled Pearse as Saturnalia 1.14 and separately linked 1.12.35 at LacusCurtius in HTML and README. MINOR repair recorded here only; no page correction line. - [fixed] X6: Restored the exact common-year qualification inside the HTML quotation. Added the dated correction in HTML and markdown. The README already marks its quotation's omission with an ellipsis. - [fixed] X7: The separate 1.12.35 link in HTML and README supplies the Senate decree while keeping Pearse for 1.14. Same MINOR repair as X5; no page correction line.