# Signs in the sky in the Romanian chronicles, checked one by one The chroniclers dated the sky well. They just did not always write it down. Of 22 notes of eclipses dated to the day in the chronicles of Moldavia, Wallachia and the Transylvanian Saxons, 13 fall exactly on the day written and one a day early; with the same one-day tolerance, a date picked at random would hit one in 88. Yet of 10 great eclipses that passed high over Brașov between 1520 and 1800, the Transylvanian texts note 4; they are silent on the other 6, though for 5 of these the same texts record other events within a month either side. ## How to check a sign from three hundred years ago The chronicles of Moldavia and Wallachia and the annals of the Saxons of Brașov and Sibiu keep 283 notes of things seen in the sky: eclipses, comets, red night skies, double suns. Two compilers searched the old editions independently, without seeing each other's lists: they found 289 and 285, 271 in common; after reconciliation 283 remain. Each date was then converted (years 'from the creation of the world' begin on 1 September; Transylvania moved to the new calendar in December 1590, Moldavia and Wallachia did not). Eclipses were set beside the sky computed for that place and day, twice: with NASA's elements and with the JPL DE422 ephemeris, whose solar-eclipse magnitudes differ by at most 0.013 of the diameter. Comets and auroras were compared with the catalogues described in the method; the other signs got no verdict. To know what a match is worth, every date was moved at random by 5 to 50 years, a thousand times. With a one-day tolerance, an eclipse dated to the day 'matches' by luck about once in 88. > 23. Sept. A media 10. usque ad 12. meridianam eclipsis solaris Coronae visa est. > — Martin Ziegler, Calendarium historicum, Quellen zur Geschichte der Stadt Kronstadt IV (1903), p. 533 The day is exact, and the approximate hours fall inside the computed eclipse: start 10:12, end 12:59 sundial time, 90.7% of the disc covered at 11:35. ## When the diarist could no longer read Johannes Lutsch, a Saxon of Sibiu, was an envoy in Constantinople in the spring of 1661. On 30 March he wrote in his diary that at noon it grew so dark that printed letters could not be read. > zwischen 12, und 1 Uhr zu Mittag war ein gross Sonnenfinsternuss dermassen, dass es so dunkel ward, dass man die gedruckte Schrifft nicht lesen konnt, und währet biss man ein Hüner Eychen hätte können braten. > — Johannes Lutsch, jurnal / diary, Kemény, Deutsche Fundgruben I (1839), p. 334–335 True: at Constantinople the eclipse was total, greatest at 12:08 sundial time. Totality lasted 4 min 6 s. ## The right day, the wrong body Martin Ziegler, once rector of the Brașov school and by then pastor at Bod near Brașov, noted on 3 May 1715 an eclipse of the Moon 'visible here'. There was no lunar eclipse that day. There was a solar one: at Brașov the Moon covered 58% of the disc, near noon, with the Sun high. A slip of this kind, the exact day with the other body, turns up at a random date about once in a hundred (1%). Pavel Binder, who collected the note in 1980, read '7 May'; the printed page says 3. > 1715, 3. Maii. Contigit eclipsis lunaris apud nos visibilis. > — Martin Ziegler, Calendarium historicum, Quellen IV (1903), p. 533 The exact day, but the eclipse was of the Sun, not the Moon. ## Three eclipses in a year, one a month too late The Jesuits of Brașov listed three eclipses in their report for 1750: a solar one on 8 January and two total lunar ones, on 19 July and 13 December. The first and last are correct to the day. Nothing happened on 19 July: the total lunar eclipse was on 19 June, the same day of the previous month (at a random date: 2.6%). They also write that first-magnitude stars were seen during the January eclipse, though the Sun was only 53% covered by area. The one bright 'star' in the sky then was Venus, near its brightest, which can be seen by day. And the hour written for the lunar eclipse, 'about the eleventh hour of the night', fits the real eclipse of 19 June: the month is wrong, not the observation. > Altera ecclipsis totalis accidit in luna 19. Iulii circa undecimam noctis > — Historica relatio, misiunea iezuită din Brașov / the Jesuit mission at Brașov, Quellen VII (1918), p. 567 A real eclipse, a month earlier: 19 June. ## Stars by day, dated a week out A note in the Latin annals of Transylvania (Chronicon Fuchsio-Lupino-Oltardinum, in the notes of the 1847 edition) enters under 16 April 1567 a 'horrible' eclipse with stars in the sky. The eclipse was real and almost total over the town (99.4% of the disc), but on 9 April, a week earlier, also a Wednesday. Likewise Laurentius Kusch puts the eclipse of 1654 on 15 August, three days after the real one, while the Sighișoara chronicle has it on 12 August, exactly. Of the 8 notes that fail the test, four have an eclipse of the same body a few days or a month away, two one of the other body (1453 and 1715), one an eclipse on the written day but below the test's threshold (1608, at Sibiu), and one no eclipse at all (the darkness of 1605/1606). > Horrenda Ecclipsis d. 16. April. accidit, ut sol non appareret, et stellae conspicerentur. > — Chronicon Fuchsio-Lupino-Oltardinum, ed. Trausch, 1847, I, p. 66, notă / note A real eclipse, but on 9 April: a week earlier. ## The eclipse that moved to the fall of Constantinople One branch of the manuscripts of the Cantacuzino chronicle, printed in the 1960 edition from manuscript O, copied around 1800, adds that on the day the Turks took Constantinople, Tuesday 29 May 1453, 'there was also an eclipse of the sun that day'. There was no solar eclipse then. Exactly one week earlier, also a Tuesday, the Moon rose eclipsed over the besieged city: the partial lunar eclipse of 22 May. A match 'exactly one week off' turns up at a random date in 1.5% of cases. The sentence belongs to this later redaction, not to the chronicle's base text. > Făcutu-s-au într-ace zi și eclipsis soarelui, ca să cunoască toată lumea că s-au dat împărătească cetate în miinile păginilor. > — Letopisețul Cantacuzinesc, manuscrisul O / manuscript O, ed. Grecescu & Simonescu, 1960, anexa, p. 201 No solar eclipse. The lunar eclipse was on 22 May, a week before. ## Moldavia: the right sky, the wrong year Nicolae Costin describes a star with a tail that appeared on 10 December (old style), 'so long it took in half the sky', starting low 'towards Hungary', rising day by day towards the north, and lasting 'seven weeks and four days, until the first of February'. The description fits the Great Comet of 1680: computed from its orbit, on 20 December (old style), ten days after Costin's date, the head stood 14° above Iași's western horizon at nightfall, its tail rising from the west towards the north. The year is wrong: Costin writes 7188, that is, December 1679 (the edition prints 'from Christ 1689' beside it, and 1679 a paragraph later). It should be 7189. Ureche likewise puts the great comet of November 1577 a year early (in his year 7085). That is what would come out if a year 'from Christ' had been converted by adding 5508 for the autumn months too, where 5509 was due: a hypothesis resting on two cases, not a fact. Neculce puts the 1699 eclipse on Tuesday 12 September, the day Costin gives for the handover of Kamianets. Costin puts it the next day, Wednesday the 13th: Costin is right. > La cursul anilor 7188, iară de la Hristos 1689 [sic], în luna lui Decemvrie 10 zile ivitu-s-au pre cer o stea cu coadă, căreia îi zic latinii cometa, adecă mătură. > — Nicolae Costin, Letopisețul Țării Moldovei, transcris în ortografia de azi / modern spelling; ed. Kogălniceanu, 1872, II, p. 20 The description fits the comet of 1680, but the written year corresponds to 1679. ## What they did not write: 25 July 1748 Around one o'clock on 25 July 1748 the Moon covered 88% of the Sun's disc over Brașov, with the Sun 60° high. The town's annals for that year survive: a turner who collapsed and died in church at Pentecost, a woman made to wear the shame stone, a regiment leaving in August. Other Brașov texts note things on 24 June and 3 July, and on 29 July, four days after the eclipse, 'again many locusts'. Of the sky, nothing. Between 1520 and 1800, 10 great eclipses (over 80% of the Sun's diameter, with the Sun at least 10° high) passed over Brașov. The Transylvanian texts note 4: 1567 (a week off), 1590, 1654 and 1699. They are silent on 1544, 1605, 1661, 1666, 1748 and 1793; for 5 of these, the same texts record other events within a month either side. The 1661 eclipse was noted only by a Sibiu Saxon then in Constantinople. A surviving text's silence does not mean nobody saw the eclipse. > 1748 den 29. Juli fanden sich abermal viele Heuschrecken auf unsern … > — J. Teutsch, la patru zile după eclipsă / four days after the eclipse, Quellen IV (1903), leaf 200 A great eclipse, unrecorded. ## Every sign on one line Above, every note in the chronicles at its year, coloured by verdict. Below are the solar eclipses computed to have been visible from Brașov with more than half the Sun's diameter covered; larger circles mean more of it covered. Threads tie a note to the real eclipse where the written date is off. ## What a match is worth A match proves nothing if it would have happened anyway. The table sets beside each kind of sign how often the real sky matches and how often it would match a date picked at random. Comets and auroras dated only to the year say almost nothing: nearly half of all years had a comet somewhere. Halos, double suns, falling fires and undescribed 'signs' (111 notes) have no independent record to compare with. They are counted and shown, but get no verdict. ## One year, one town: what the chroniclers wrote and what the sky held Pick a year between 1450 and 1800 and a town. The eclipses are computed now, for that town, from NASA's elements; the notes come from all the chronicles, wherever they were written. ## The method and its limits The chronicles. The old public-domain editions: Ioan Bogdan, The old Moldavian chronicles (1891); Mihail Kogălniceanu, Chronicles of Romania (1872–1874: Ureche, Miron and Nicolae Costin, Neculce, Muste, Enache Kogălniceanu and others); Magazinu istoricu pentru Dacia, vol. V (1847); Quellen zur Geschichte der Stadt Kronstadt IV–VII (1903–1918); Chronicon Fuchsio-Lupino-Oltardinum (1847–1848); Kemény, Deutsche Fundgruben I–II (1839–1840). The Cantacuzino chronicle was read in the 1960 critical edition. Not readable here: Magazinu istoricu vols. I–IV (printed in the transitional alphabet, the scans' text is illegible), so Radu Popescu, Radu Greceanu and Constantin Căpitanul are missing. Two independently compiled lists. Two compilers (Claude agents and Codex) extracted the notes separately: 289 and 285, 271 in common (89% of the union). All 20 signs that Moesz (1911) and Binder (1980) cite from the same volumes are in the union. Dates were then read on the page and transcribed without conversion; one program does all conversions. Two computations. Solar eclipses: Besselian elements from NASA's Five Millennium Canon (Espenak and Meeus) and the JPL DE422 ephemeris through Skyfield, each with its own ΔT; they differ by at most 0.013 of the diameter. Lunar eclipses: NASA's Canon and Skyfield. Comets: Pingré, Cométographie (1783–1784). Auroras: the Křivský and Pejml catalogue (1988), without the Hungarian records when the sign comes from Transylvania, so no chronicle confirms itself. The rules were fixed before comparing, tolerances and random-date test included. What was added after a first look at results is marked as such: the 'moved by days, a week or a month' readings (five readings tried on each missed eclipse; by chance about one hit per three sets of eight would be expected, six came out), the hours test, and the list of unrecorded great eclipses, checked by reading. The hours the chroniclers wrote are generally within an hour of the computed ones; this suggests first-hand observation but does not prove it. What the page does not show. The drawn sky is lit by the uncovered part of the Sun; light from beyond the Moon's shadow is not modelled, so near totality the real horizon is somewhat brighter; at totality the page gives the sky, as a stand-in, a ten-thousandth of normal daylight (the page's choice, not a measurement). For the great eclipses not recorded, a surviving text's silence does not mean nobody saw them. ## Sources - I. Bogdan, Vechile cronice moldovenesci până la Urechia (1891) https://archive.org/details/ioan-bogdan-vechile-cronice-moldovenesci-pana-la-urechia-texte-slave-cu-studiu-traduceri-si-note - M. Kogălniceanu, Cronicele României, 2nd ed. (1872–1874) https://archive.org/details/croniceleromanie13koga - Magazinu istoricu pentru Dacia, vol. V (1847) https://archive.org/details/magazinuistoric03unkngoog - Quellen zur Geschichte der Stadt Kronstadt, IV–VII (1903–1918) http://dspace.bcucluj.ro/handle/123456789/178034 - Chronicon Fuchsio-Lupino-Oltardinum, ed. Trausch (1847–1848) http://dspace.bcucluj.ro/handle/123456789/152629 - J. Kemény, Deutsche Fundgruben der Geschichte Siebenbürgens I–II (1839–1840) https://archive.org/details/deutschefundgrub01kem - Istoria Țării Românești 1290–1690. Letopisețul Cantacuzinesc, ed. C. Grecescu and D. Simonescu (1960) https://archive.org/details/LetopisetulCantacuzinesc - F. Espenak and J. Meeus, Five Millennium Canon of Solar / Lunar Eclipses (NASA) https://eclipse.gsfc.nasa.gov/SEpubs/5MCSE.html - JPL DE422 planetary ephemeris https://ssd.jpl.nasa.gov/planets/eph_export.html - A. G. Pingré, Cométographie (1783–1784) https://archive.org/details/comtographieou001ping - L. Křivský and K. Pejml, Solar activity, aurorae and climate in Central Europe (1988), NOAA catalogue https://www.ngdc.noaa.gov/stp/space-weather/aurora-airglow/aurorae.dat - M. Stangl and U. Foelsche, Aurora observations from the Principality of Transylvania, Solar Physics 296 (2021) https://doi.org/10.1007/s11207-021-01811-7 - H. Hayakawa et al., A great space weather event in February 1730, A&A 616 (2018) https://doi.org/10.1051/0004-6361/201832735 - The drawn atmosphere: the model of 'The blue after sunset' (Bruneton 2017) https://mariuscomper.uk/albastrul-de-dupa-apus/