The Solar Eclipse of the Xia Dynasty: A Review
aa r X i v : . [ phy s i c s . h i s t - ph ] J un G ERMAN V ERSION :see Habilitation at the University of Heidelberg arXiv: 2006.04674 [physics:hist-ph] v2: 28 June 2020
The Solar Eclipse of the Xia Dynasty: A Review
Emil Khalisi [email protected]
Abstract
We present a review on the renowned solar eclipse in the Chinese book
Shujing describing the oldest accountof this kind. After balancing the arguments on the time, place, and the celestial stage, we offer a new scenario.The path of totality on 15 September 1903 BCE traversed Anyi, the assumed capital of the Xia dynasty.The date matches remarkably well the chronological order of two other incidents of mythological rank: theclosest-ever agglomeration of the naked-eye planets and the Great Flood under Emperor Yu. Still, this eclipsedoes not remove all questions about the historical circumstances given in the account.
Keywords:
Solar eclipse, Astronomical dating, Chronology, Xia dynasty, Ancient China
The solar eclipse in the reign of the Xia Emperor ZhongKang is one of the best known among the historical ac-counts on eclipses. The event is associated with the dreadfulfate of two astronomers, named He and Ho, who were notprepared for it and therefore punished. This oldest accountknown, held in the classical work
Shujing (also
Shu Ching or Shoo King ), surpasses any other by more than 800 years.A lot of information about this eclipse is rather legendarythan steady, as will be shown.To understand the connection between the mytholo-gical character, its fame, and the reality, one has to tracedown the history of the sources as well as the political im-portance of astronomy in the days of ancient China. Thisarticle tries to review the passed down facts. First, webriefly examine the origins of the
Shujing and the derivativedescriptions and check for their conclusiveness. We quotethe relevant paragraphs from these works. After summar-ising previous analyses, we compare the proposed dates,and finally suggest a new solution that would be consistentwith two other incidents of that obscure era. In conclusion,we wish to add the new date to the evidence in favour of thehistoricity of delineated events, although the account itselfis of doubtful nature.An essential quantity for studying historical eclipses isthe “clock error”, ∆ T . It denotes the deceleration parameterof the earth’s rotation: the difference between a theoretic-ally uniform time scale (Ephemeris Time) and the presenttime (Universal Time). This issue will not be covered here,but we touch on the problem at the very end pointing outthat our result perfectly fits the average value of the extra-polated ∆ T .Dates will be given historically, i.e. omitting the yearzero. For example, the year “1903 BCE” is “-1902” of theastronomers. The Xia dynasty is the first kingdom that controlled a ma-jor domain of tribes in China. According to traditional textsthere existed “Five proto-Emperors” with some kind of god-like status before the Xia. Some knowledge about these le-gendary Emperors helps to understand the events precedingthe solar eclipse.
Almost all we know about the solar eclipse at the dawn ofChinese history goes back to the
Shujing , the “Book of His-torical Documents”. Its origin is particularly complex. Ap-parently written by various authors, it is attributed to thephilosopher Confucius (551–479 BCE). The book containsspeeches, royal decrees, proclamations, and appointmentsof high officials since the foundation of the kingdom. Em-bedded into an elegant language, many important events areconveyed there. Starting with those pre-dynastic emperors,the oldest documents are estimated at 2400 BCE, the latestat the 7th century BCE [7]. The collection contains no con-tinuous historical narrative and no attempt at chronology.The original of the
Shujing was destroyed in the burn-ing of books in 213 BCE, together with many other works.This event represents a pivotal moment in the history ofChina, as the monarch of the short-lived Qin dynasty per-formed various acts against Confucianism and other beliefs.He wanted to reform the state and re-write history fromscratch. The destructive operation was to leave behind agreat impact on the society and culture. In spite of the im-mense loss of precious historical treasures, a few changesled to some positive aftermath in the following Han dyn-asty. However, the burning of the books is the main causefor our lack of knowledge about the very early times.Attempts were made to recover the book. Fragments . Khalisi (2020): The Solar Eclipse of the Xia Dynasty: A Review and a table of contents escaped the destruction. An incom-plete copy of the Shujing could be found about 25 yearslater. It is said to have consisted of 28 or 29 out of theoriginal 100 chapters, but there was no mention of an ec-lipse [1]. Then, a ruler in the 4th century AD ordered torestore the book, i.e. more than a half thousand years afterthe loss. Some citations were gathered from other books,which one could get hold of, other parts were possibly re-plenished. No matter whether or not the restoration couldsucceed, it seems obvious that gaps would be inevitable. Wemight call it a forgery [7]. But the biggest defect concerns,from our present point of view, that the unknown restorerdid not specify when he was filling in or when conjecturinga possible story, and even not which sources he followed. Itseems that he just headed for the result rather than cared forthe route towards it.Most scholars of the 19th century put forward that thedocuments were selected by Confucius not for their histor-ical importance but for teaching morality [23, 22, 3]. Thelapse of the astronomers He and Ho did not consist of themissing prediction of an eclipse but the disregard of theirduty. The royal officers were to superintend the customaryrites. These rites were to apply the royal regulations likecommissioning archers, beating of drums, changing clothes,igniting incenses, and something like that. The emperorshould perform ceremonies and carry out certain acts to“prevent the world from destruction”. The Chinese beliefwas that a dragon or monster would approach the luminaryand threatened to devour it. The people were to dispel thedragon by making noise with all kinds of equipment. Fortu-nately, they succeeded each time.The existence of royal instructions for solar eclipsespresupposes some knowledge about the phenomenon. Peoplecould not forecast it, but they knew how to cope with it.There were persons in charge and they had to administerthe procedure. Though nobody could predict dates, peoplecould react accordingly as soon as the “dragon” was advan-cing.Besides the
Shujing there are two more passages fromother sources referring to most likely the same eclipse. Thesecond one emanates from the
Annals of the Bamboo Books .Before the invention of paper in the first century BCE, bam-boo was a common writing medium, and strips of typicallength of 50 cm were tied together to form a book. Thischronicle escaped the burning of books just by chance, be-cause it was entombed with the relicts of a king who haddied in 296 BCE. It was discovered, together with otherscripts, by tomb raiders in 281 AD [20]. Some parts ofthe bamboo strips were destroyed but the remaining wererestored and copied several times. The
Bamboo Annals also report about those pre-dynastic emperors as well asthe dawn of the Chinese culture. The history ends with theaforementioned king who kept it in his tomb.The third text is by the historian Zuo Qiuming (ca. 556–451 BCE). He lived in the state of Lu which was the homeof Confucius. Some historians say that he was a student ofthe latter and did not make as large an appearance as otherstudents, thus, he is less prominent. Other historians just
Figure 1: Approximate territory of the Xia Dynasty and theinundated areas [30]. tell that he was contemporary to Confucius. Zuo Qiumingis best known for his commentary on the ancient chronicle
Chunqiu or “Spring and Autumn Annals” — not to be con-fused with the
Bamboo Annals above. Written in narrativestyle, it covers the period from 722 to 468 BCE and focusesmainly on political, diplomatic, and military affairs fromthat era.
The introducing chapter of the
Shujing begins with a glor-ious regent Yao who is hardly datable. Some put him inthe middle of the 24th century BCE, others to the end ofthe 22nd century BCE. For example, the translator of the
Shujing , James Legge (1815–1897), fixed his accessionto 2357 BCE, while the astronomer Ludwig Ideler (1766–1846) did that to 2163 BCE [13].Yao was one of those Five Great Emperors before thefirst hereditary dynasties. According to the
Bamboo Annals he died at the age of 119. Before him there were othersupremes with reigns of more than 10,000 years. Suchstatements deprive of dating in general. Any informationfrom the early times remains extremely fuzzy till ≈ Shujing is that Yaoordered to determine the cardinal points of the sky and foun-ded the calendar “at the beginning of time”. He was ad-vised by two astronomers named He and Ho. They wereto observe carefully each celestial part represented by thesolstices and the equinoxes, respectively [22].During the reign of Yao there would have been a tre-mendous flood, also of mythological rank (Figure 1). Thepassage referring to it appears later on the documents, anda similar account was given by other historians between the4th and 1st century BCE [18]. If one gives credence to therecords, the deluge was the most severe since the ice age.Geologists assume a massive earthquake in Eastern Tibetthat led to landslides, dammed lakes as well as redirections . Khalisi (2020): The Solar Eclipse of the Xia Dynasty: A Review Table 1: Proposals for the reign of the Xia dynasty.
Time span [BCE] Method of dating Ref.2205 – 1766 Traditional chronology [18]2200 – 1675 Timeline of rulers (Wikipedia) [30]2183 – 1751 Solar activity / method unclear (1980) [27]2100 – 1800 Radiocarbon method (1992) [5]2070 – 1600 Xia-Shang-Zhou-Project (1996–2000) [14]1989 – 1558 Annals of the Bamboo Books ( ≈ -300) [13]1953 – 1550 Planetary grouping + average regency [18]1900 – 1550 Geologic stratigraphy (2016) [32]of rivers. It would have modified the landscape and alteredthe agricultural conditions. The disaster has been dated at1922 ±
28 (1 σ ) BCE utilising the radiocarbon method onskeletons of three cave dwelling human victims [32]. The95% confidence interval ranges from 1976 to 1882 BCE.This geologic study provoked controversy among schol-ars. We will not re-discuss the pros and cons here, see [33]and references therein. Another work argues that the samesediments, that have been used in that analysis, indicaterather two separate landslides having occurred at 8,300and 6,300 BCE, respectively [34]. The two dammed lakesgradually shallowed and then disappeared, but both pre-ceded the Great Flood of the Xia by 1400 years, at least.The geomorphic events would be unrelated to the historicalevents in the writings.The legends say that the Great Flood lasted for two gen-erations, putatively two decades, at least, till Yao’s secondsuccessor, the engineer Yu, attained a solution to the prob-lem. After “taming the waters”, he was conventionalised asa hero, then ennobled as emperor and founder of the firstdynasty, the Xia. He was given the epithet “the Great”.Furthermore, the calendar ought to have started with him(again?) upon a celestial sign which will be outlined below.Yu is said to have ascended 100 years after Yao. This contra-dicts the statement on the inundation of “two decades”, butany information might already be rooted in later legends.The Flood marks a central element for the first dynastyof kings/emperors. Manifold efforts were made to placeit at the start of the Chinese chronology. Some have givenspecific years, while others consider the problem intractable(Table 1). An archaeological historicity can only be verified for thesecond dynasty, the Shang, which started in the mid-secondmillennium BCE. There are indications that Xia and Shangwere coexisting and interacting spheres of influence vyingfor supremacy [5, 18]. Even some coeval intervals in theking lineages of the two dynasties have aroused suspicion.A long-standing debate is related to whether or notthe Xia was identical with the so-called Erlitou culture, anurban society, that existed in the Yellow River valley atthe same time. Guided by some geography texts, excava-tions were performed at a site Yangcheng in Honan. The place would be consistent with Anyi, the probable resid-ence of the Xia kings, though there are different locationssurmised. Pottery was recovered, ceramic vessels, andpieces of bronze. One sample has a tree-ring-corrected ageof 1900 ±
70 BCE [18]. However, this gives no proof forthe Xia, since the affiliation of the archaeological samplesremains unclear. Because of this deficiency, the time forthis dynasty is given very, very roughly depending on themethod of analysis, see Table 1. Its starting point may fallat some time within the interval between 2200 and 1900BCE. The endpoint of the Xia is not known for the samereasons.Seventeen monarchs are known from the Xia, but thelengths of their regencies are rather estimated. The ancienthistorian Sima Qian (ca. 145–86 BCE) gave a relative or-der for their sequence, but even the names have been as-signed posthumously, for they are not well identified. Mostevents from those times originate from texts compiled dur-ing the Eastern Zhou Dynasty (from 771 BCE onwards).How they were transmitted over many centuries remains un-known, too. Actually, the whole existence of the Xia restson legendary stories.
The close approach of three or more planets is usually amagnificent, eye-catching sight in the sky. In the history ofmankind there were social turmoils at such occasions. Thisissue is widely ramified and has been discussed extensivelyunder astronomical as well as historical aspects elsewhere.Almost every culture considered tight configurations ofplanets as the moment of creation. It was linked to the be-ginning of time and set equal to the onset of the particularcalendar. In Babylonia, India, Greece, and in the MayanEmpire astronomers computed great cycles for the cosmosbased on the re-occurrence of special planetary meetings.First an apocalypse was prophesied, often combined with adeluge, thereafter a re-creation with a “new era”. Ancientscholars believed that any cosmic cycle would have startedfrom a certain point in the past: the day, the lunation, theyear, and the course of the planets.An overwhelming clustering of planets did happen inspring of 1953 BCE. All five classical planets gathered atdawn with the moon and the sun joining the parade. Thiswould have triggered the Chinese calender. The extraordin- . Khalisi (2020): The Solar Eclipse of the Xia Dynasty: A Review Figure 2: Closest-ever agglomeration of the naked-eye planets on 25 February 1953 BCE, 23:36 UT, at sunrise in Anyi. ary grouping is accompanied by a remark in the work by thelibrarian Liu Xiang (77–6 BCE) [19]:The original
Zhuanxu calendar began . . . oncyclic day 6, mon 51, year 51 (modulo 60) atthe start of spring when the sun, moon and 5planets met at
Yingshi , 5 ◦ .The date is correlated with 5 March 1953 BCE. The po-sitions of the planets reached their smallest separation sevendays prior to that, namely on 26 February, with a minimumdistance of 4.3 ◦ (Figure 2). That value was never beat inthe reproducible history such that it is regarded as the nar-rowest grouping ever. Robert Weitzel was the first to drawattention to the scene as an astronomical curiosity, but hewas unaware of the historical record by the Chinese [28].On 2 March, the planets were visited by the waning moon,and a few days later the new month began with the lunarcrescent in the evening time.The massing of planets took place in the constellationof “Yingshi”, indeed: It is allotted to Pegasus near the bor-der of Pisces and Aquarius not far from the equinox. Pe-gasus was rising, but the planets could not be seen well dueto the proximity of the sun (too small difference in height).Sometimes the above quotation is assigned to the myth-ological proto-Emperor Zhuanxu who reigned two positionsbefore Yao. Texts depict Zhuanxu as a god who separatedHeaven from Earth and arranged the positions of the sunand stars. At the age of twenty he became sovereign, goingon to rule for fabulous 78 years similar to his predecessorsand successors. Various authors place Zhuanxu somewhereat ≈ ◦ . Historically, the Xia dyn-asty was replaced by the second, the Shang. The new rulerclaimed this clustering as his “legitimation” for the com-mand. At that occasion the calendar was reformed deploy-ing a 60-day-cycle. This cycle gives a special designationto each day. The mechanism is up and running strictly andcontinuously to our days without any disruptions.Then, in late May 1059 BCE, the naked-eye planets ap-peared for a rendezvous (6.5 ◦ ) once more, and the freshlyestablished Zhou dynasty declared it to its “Heavenly Man-date” [20]. Again, the calendar was manipulated and thensuspended by a later dynasty. The Han, that just conqueredpower in 206 BCE, made use of another gathering of planetsin May 205 BCE, though it was less impressive ( ≈ ◦ ). . Khalisi (2020): The Solar Eclipse of the Xia Dynasty: A Review All kinds of calendar reforms have been, at any time, anappropriate remedy for practising political control. Manysovereigns tried to customise the history of the country totheir own needs. All these “conjunctions” can hardly beused for dating purposes, for they have been reckoned inlater times to suit the astrological influence “justifying”their policy [11].
On the whole, the Chinese history consists of many divi-sions as well as rivalling monarchs. The times of their ruler-ship were differently reckoned depending on the source ofinformation. Their flourishing is of great importance forboth historians and astronomers.For the purpose of a binding chronology the so-calledXia-Shang-Zhou Project was born in 1996 [14]. It collatedmore than 200 scientists from various fields of work amongthem archaeologists, astronomers, historians, and palaeo-graphs. The objective was to explore the period of actionfor the first three dynasties. The experts should reconcilefixing points for the old data with our modern chronology.The Project was carried out by 44 working groups, twelveof which used mainly astronomical evidence. As for theXia, a decision was made with the commencement of thisdynasty placed at approximately 2070 BCE (Table 1).The Western world looks more critical at this nationalproject. Sinologists consider many conclusions as inad-equate, see e.g. [11]. New evidence has it that the new dat-ing by the Chronology Project also turns out to be flawed.Recent archaeological discoveries like the sculptural bronzeartifacts from the Meixian County undermine almost everyone its dates [24]. So, the controversy surrounding the solareclipse of our interest during the Xia dynasty is not settledyet.
The next reference after Yao to an astronomical incident ap-pears in the
Shujing at the time of the fourth Xia-EmperorZhong Kang. In his first year occurred that solar eclipsethat became so popular among the legends. We quote fourversions.
In the order of progression, the preface to the Book III ofthe
Shujing points to “The Punitive Expedition of Yin”. Theprince of Yin was sent out to punish He and Ho because theyallowed the days to get into confusion [13, p3]:(1) When Zhong Kang commenced his reignover all within the four seas, the prince of Yinwas commissioned to take charge of the im-perial armies. At this time He and Ho had neg-lected the duties of their office, and were sunkin their private cities, and the prince of Yin re-ceived the imperial charge to go and punishthem. Then, Book IV describes the mission of the prince onbehalf of the emperor in more detail. In the verses 2 and3 the prince makes a military announcement. He addresseshis forces on the objective of the expedition and says thatthe stability shall be restored in the land. So far, there is nomention of an eclipse. The reader does not come to knowwho the two persons were, nor their duties, nor how theycommitted their crime. The laws are put forward and thenit is said [13, p162]:(4) Now here are He and Ho. They haveentirely subverted their virtue, and are sunkand lost in wine. They have violated the du-ties of their office, and left their posts. Theyhave been the first to allow the regulations ofheaven to get into disorder, putting far fromthem their proper business. On the first dayof the last month of autumn, the sun and themoon did not meet harmoniously in
Fang .The blind musicians beat their drums; theinferior officers and common people bustledand ran about. He and Ho, however, as if theywere mere personators of the dead in theiroffices, heard nothing and knew nothing; —so stupidly they went astray from their dutyin the matter of the heavenly appearances,and rendering themselves liable to the deathappointed by the former kings. The statutesof government say, “When they anticipate thetime, let them be put to death without mercy;when they are behind the time, let them beput to death without mercy.”The concluding three verses of the Book IV contain in-vocations to the enemies to surrender, since the legitimationfor the punishment was ordered by the heavens. — Thereader is informed about a turmoil in the country.
The second passage is found in the
Annals of the BambooBooks . The incident is recorded as follows [13, p119]:(1) In his first year, which was ke-ch’wo (26thof the year cycle = 1951 BCE), when the em-peror came to the throne, he dwelt in
Chin-sin .(2) In his 5th year in the autumn, in the 9thmonth, on the day kang-siu (47th of the daycycle), which was the first day of the month,there was an eclipse of the Sun, when heordered the prince of Yin to lead the imperialforces to punish He and Ho.The calendric and territorial information will be ana-lysed in the next section. About the capital, which is saidto be
Chin-sin , the translator, James Legge [13], remarksin a footnote: “The site of
Chin-sin is not well ascertained.The dictionary places it in the district of Wei, department ofShan-tung. Others — more correctly, I think, — refer it tothe district of Kung, department of Ho-nan.” . Khalisi (2020): The Solar Eclipse of the Xia Dynasty: A Review Deviant from the
Shujing it is mentioned here that theeclipse occurred in the fifth year of the Emperor. Whenreading the text independently, the crime of He and Ho be-comes not clear. It is not explained what the eclipse has todo with them. Their fault could lay some time before it. Inother words: there are two people being punished on a daywhen, by chance, a solar eclipse happened? [6]Moreover, the name of the punishing prince is not given,again. “Yin” is — as in the
Shujing above — the regionwhere he lived or descended from. The modern encyclopae-dia give a mountains to the North of China.
The confusion widens as one takes notice of the third text byZuo Qiuming. He writes in his commentary to the “Springand Autumn Annals”,
Chunqiu , about an eclipse that wasto take place in the 6th month, in summer. At the decis-ive point a discussion is going on between a priest and ahistorian. The historian can be set equal to the astronomeror astrologer who takes care of the chronology and calen-dar. The two converse about the rites to be performed in thecase of the eclipse: to beat the drums, wearing of clothes,and presenting gifts. While quarrelling over what protocolshould be followed, the historian/astronomer intercalatesand recalls a former eclipse long ago [7, 27]:The Sun already passed the equinox but hasnot arrived at the solstice. When any calam-ity happens to the three celestial bodies (i.e.Sun, Moon, and planets), the various officialsput off their elegant robes, the king does nothave his meal (table) fully spread, and with-draws from his principle chamber, till the timeis past. (. . . ) This is what is written in the
Xia Shu : “The Sun and Moon could not liveharmoniously in their place, the blind beat thedrum, low rank officials mounted the horses,and people ran up in haste.” That is said ofthe first day of this month — it was in the 4thmonth of Xia, which is called the first monthof summer.The words within the quotation marks used by the royalastronomer are identical with those in the
Shujing . Thereseems no doubt that they refer to the same solar eclipse ofXia. According to Schlegel & Kühnert, the word “equinox”alludes to the spring, and the instructions would only ap-ply to eclipses in summer [23]. The circumstances wouldresemble those old days of Xia.Though not dealing directly with the story of He andHo, the text throws instructive light on the eclipse customsin ancient China. Finally, we get the impression that theaftermath of the solar eclipse encompassing He and Ho re-mained in mind for very long. Now we have three quota-tions at hand bearing reference to the same incident, butthey are ambiguous.
In modern times, a completely different variant emerged.The misdoing of the two astronomers would be contrary:they foretold a solar eclipse that did not occur. The conceptis based on the idea that the monarch possessed the powerof darkening the sun whenever he wanted. Actually, he re-lied on a man skilled in foretelling eclipses. The two astro-nomers got secret information about the next date by a sub-terfuge and wanted to benefit from it over the emperor. Theemperor gathered the public upon the prognostication in or-der to distinguish himself by the miracle of commanding thesun. But the foreteller intentionally included a wrong date,and the occultation failed to appear. After the proclamationhad gone wrong, the monarch was embarrassed so muchthat the doom of the sinners was the same.This alternative entertains with a love story and in-trigues. It has its roots in the novel “Sonne, Mond undSterne” by A.G. Miller, which is a pseudonym of theGerman-Austrian writers Marie Louise Fischer (1922–2005) and Hans Gustl Kernmayr (1900–1977). Althoughthe book is categorised as non-fictional, the storyline suitsrather a thrilling TV movie than sound history [10]. Thisversion can be rejected, for in those times there was defin-itely no-one to presage a solar eclipse.Within the alleged ≈
180 years between the first Em-peror Yu and the fourth Zhong Kang about 15 solar eclipsescould have been visible, mostly partial ones. This is muchto little to discover an astronomical cycle, even if all ofthem would have been observed [23]. Another constraintconcerns the fact that there was no eye protection for mon-itoring the sun permanently. It was not useful, anyway, andthus only those eclipses would be seen that are almost total,or close to the horizon at sunrise and sunset. Many partialobscurations would have passed unnoticed. Therefore, thetwo unfortunate astronomers could never have met the re-quirements of their office.
Countless attempts were made to identify the mysterious ec-lipse in spite of objections by historians. From the first text,the
Shujing , only one line has an astronomical relevance:The Sun and the Moon did not meet harmoniously in
Fang on the first day of the last month in autumn.One might think that it will be simple to find a suit-able event having the location and date at hand, but, unfor-tunately, it is not. The main difficulty concerns the interpret-ation of any information: the specification of the season isdisputed as well as the celestial meaning of “Fang”.
Usually,
Fang is assigned to an area of about 5 ◦ between β , δ , π and σ Scorpii, a region in the “pincer” westwardof Antares. A few stars from Libra and Ophiuchus wouldbelong to it, too [31]. The Sun needs 5 days to cross thispart of the sky. According to the Xia calendar, it is the ninth . Khalisi (2020): The Solar Eclipse of the Xia Dynasty: A Review month, the month after that which contains the autumnalequinox.Other commentators consider Fang as undeterminedbut understand it more general as a “planetary mansion” or“domicile”, i.e. those 28 Chinese lodges along the eclipticwhich can be visited by the Moon [23, 2]. Each of the con-stellations has a different extent: the smallest about 1 ◦ , thelargest 31 ◦ . Thus, Fang would not be used as a fixing pointfor the position of the sun. Also, a calligraphic confusioncould not be excluded among the thousands of symbols inthe Chinese language.The statement that the sun and the moon “did not meetharmoniously” also turns out quite ambivalent. Such an ex-pression exists nowhere else in Chinese literature except forthe
Shujing . When denoting an eclipse, the common usagewas something like “the sun was devoured” or “it lost itslight” [23]. A variant of the original phrase might read “thesun and the moon could not live peacefully together in thesky” [27]. Another translation is “[they] were not in agree-ment” [31]. The crucial point applies to the Chinese symbolfor the sun: Depending on the context it could be altered in“time” or “celestial body”, or it can denote a morning hourbetween 7 and 9 a.m. [17]. This gives a different meaningto the whole line. Since most experts base their interpreta-tion on the authority of James Legge [13], the author of thispaper will stick to that. It is generally agreed upon that weare treating a solar eclipse here. The chaos among peopleand the consequences for the victims suggest that no otherphenomenon is the subject. In addition, the debate in the
Chunqiu explicitly deals with this issue.As far as the instant of the eclipse is concerned, the lastmonth of autumn is specified (
Shujing ), or the first monthof summer (
Chunqiu ), or the whole event could have beenretrieved by computation very much later. In historiographysuch acts were carried out quite often, as many examplesshow like the so-called “Star of Bethlehem” or birth datesof celebrities with political power. Some historians preferto fix the eclipse exactly to the day of the autumn equinox.In fact, it was lying in the constellation of
Fang about 2150BCE, and it would correspond to 11 October of our calendar[26].Keeping in mind that the books were written in the 5thcentury BCE, one would have to consider the luni-solar pre-cession when calculating backwards. The Chinese were notaware of the precession at the time of Confucius, though.It was discovered in the 4th century AD, more than 500years after Hipparchus [22]. Hence, there is a suspicion ofa deliberate modification by the restorers of the
Shujing assuch as they substituted “autumn” for the reading “winter”because
Fang had moved to that season. Or, taking theother perspective, autumn would be the original and
Fang was emended for agreeing with the lifetime of the writer.Most present-day analysts, however, give credence to thelast month of autumn, though there is no reason. The ques-tion about the season must remain unsettled.
The second quotation from the
Bamboo Books provides anexact day in the shape of a cyclic name: kang-siu (
Bamboo Books giving28 October 1948 BCE as the correct day for the name —3 ×
60 years later than Rothman’s result [6]. Unfortunately,there was no eclipse that day, not even a new moon. Fother-ingham put forward that the eclipse did not match the life-time of Zhong Kang and his reign would have to be shif-ted within a century or two one side or the other of 2000BCE. Instead, he emphasized the clearness with which the
Shujing makes the offence of the astronomers’ neglect ofthe calendar. It went out of order, and this became evidentthrough the occurrence of the eclipse on an unexpected day,i.e. not on the first day of the month as it should be. Thisdrew attention to their negligence and convicted them of theerror of issuing a wrong calendar [7]. They were executedfor the false reckoning and not for the eclipse. Finally, Foth-eringham concluded that the record did not permit an iden-tification of the date at all.Before that, the historian John Williams (1797–1874)used the name of the cyclic year in the previous verse, andobtained 2158 BCE as the year of the accession of the em-peror [31]. But the cycle of names was only applied toyears during the Han dynasty from ≈
200 BCE, after thecompletion of the
Bamboo Books . When looking into vari-ous encyclopaedia, the durations of rulership for all emper-ors prove contradictory. The information about the cyclicnames, whether related to the day or year, must be a laterinsertion based on a doubtful backreckoning [18].
In addition to the arguments above there are further dishar-monies: The location of the residency of the Xia dynasty isnot well known. It changed permanently with the emperor,and could be situated in Anyi or Taikang, which is 500 kmfarther to the East (Table 2).Another problem is that the text gives no hints neitherto the time of day nor to the magnitude of the eclipse: totalor a very high degree of obscuration. For example, thereare no accompanying phenomena mentioned like a suddendarkness or the visibility of stars or allusions to the corona.If the eclipse was not total, the search would expand to nu-merous partial eclipses.Also, it is not said what space of time passed after theeclipse when the “Punitive Expedition” against the astro- . Khalisi (2020): The Solar Eclipse of the Xia Dynasty: A Review Table 2: Suggested capitals for the Xia dynasty.
Place Today’s name Latitude Longitude ReferenceAn Yi Hsien Anyi 35 ◦ ′ N 110 ◦ ′ E [17, 23, 22]Tay-kang-kien Taikang 34 ◦ ′ N 114 ◦ ′ E [21, 22]Tschin-sin (in Honan) 34 ◦ ′ N 114 ◦ ′ E [23]Tschin-sin (in Shan-tung) 36 ◦ ′ N 119 ◦ ′ E [17]Yangchen Dengfeng 34 ◦ ′ N 113 ◦ ′ E [18]Yanshi Luoyang 34 ◦ ′ N 112 ◦ ′ E [5, 9]nomers was set off. The punishment could rest upon polit-ical motives. In the context of history a war against anusurper followed lasting for two years [6]. The proced-ure of punishment is not told, either. A “decapitation”, aswidely depicted, is not stated. The victims could be hangedor stoned to death or be “punished” in any other way. Thereader of the text is even not informed whether the punish-ment was actually inflicted or not. It seems enough to haveit declared that they merited death.Some modern authors stretch out the interpretation asmuch as they do not regard the names He and Ho as in-dividuals but rather the titles of officers because they arecalled this way under the pre-emperor Yao [7]. WilliamWhiston (1667–1752), a theologian and mathematicianwho succeeded his mentor Isaac Newton (1643–1727) atthe University of Cambridge, believed that both EmperorsYao (counting one astronomer named He-Ho [27]. He would be the usurper of the rebellious tribeand assembled a considerable military force to overthrowthe emperor. This discomfited Zhong Kang, so he tookadvantage of the eclipse to attack He-Ho.All these imponderables raise the thought that the sub-ject in the
Shujing is not the eclipse itself. The accountdoes not focus on the astronomical event but seems to givea moral sermon for deterrence.
The search for a solar eclipse falls into an interval of about300 years depending on the assumed commencement of theXia dynasty. Table 3 lists a selection of suggested datessorted by the year of publication.The findings scatter tremendously. The first to intro-duce the Chinese works to the Western world was the Je-suit Antoine Gaubil (1689–1759). He rejected the year of2128 BCE from medieval times, by the help of a Chineseassistant, because the eclipse would not have taken place inBeijing [8]. Instead, he introduced the date of 12 October2155 BCE.It was William Whiston who first referred to EdmondHalley’s discovery, in 1695, of the “acceleration of theMoon’s mean motion”. In 1734 Whiston pointed out thesmall quantity of obscuration ( < ∆ T ) nor the Besselian Elements were workedout. However, the date by Gaubil still prevailed for long.An improved value for ∆ T was included by Charles-Louis Largeteau (1791–1857). In 1840 he re-calculated theephemeris of the moon and found that Gaubil’s eclipse of2155 BCE occurred when China resided on the night side[23]. Also, the year 2128 BCE will have to be definitelyexcluded, in spite of a revival by John Chalmers in 1861,who was unaware of the work by Largeteau [2].In our time, the date of 22 October 2137 BCE is men-tioned most often, as suggested by Whiston (Figure 3). Thecase was examined by the astronomer and mathematicianSamuel Russell (1856–1917) in detail and seemingly dis-cussed in the most convincing way [22]. Theodor Rittervon Oppolzer (1841–1886), the time-honoured authority oneclipses, reached the same conclusion a decade earlier [17].Popular science repeats this date constantly without elucid-ating the origin of the information.However, that eclipse of 2137 BCE was not total butannular, and the magnitude in Anyi was about 0.875. Acomplete darkness was never achieved at any moment.Moreover, the date does not correspond to the (speculative)cycling day Bamboo Annals . The cyclic namewould perfectly coincide with the former 13 October 2128,which must be excluded in any case.In the year following 2137, there was another eclipsealso visible in the northern hemisphere: 11 October 2136BCE. Russell gives the time of onset at 4:30 p.m., and themagnitude in Anyi was about 0.58 [22].Including the concept of the Great Flood, which wasthe cause for the accession of the first Emperor Yu andput by archaeologists at approximately 1925 BCE, then thelate date by Kevin Pang gains a surprising attention [18].It could reduce some conflicts from the historical point ofview, though not eliminate them from the astronomical. Theagglomeration of planets cannot be integrated, for example,into Yao’s or Yu’s lifetime and his successors. And, again:the eclipse of 1876 BCE was annular, and the central trackpassed over Siberia much to the North. In Anyi, only asmall partial coverage (mag = 0.386) would be observedshortly after sunrise at 8:30 a.m. local time.The archaeoastronomer Göran Henriksson argues in fa- . Khalisi (2020): The Solar Eclipse of the Xia Dynasty: A Review Table 3: Dates of proposed solar eclipses related to He and Ho.
Date [BCE] Type Commentator (publ. year) Reference2128, Oct 13 P Yi Xing ( ≈ ≈ this work, Sec. 6 (2020) [12]vour of the eclipse of 26 October 1961 BCE because it wastotal and crossed the territory of Yin where the punishingprince had his residence [9]. Henriksson locates the regionto the Northeast of China at the estuary of the Yellow Riverabout 200 km South of Beijing. In Anyi the eclipse wouldbe seen partially with a magnitude of 0.863. The sun had adistance of 10 ◦ from π Scorpii, the central star in
Fang .After providing new limits for the Xia dynasty by theXia-Shang-Zhou Project, none of the dates suited for theeclipse except Henriksson’s. A plenty of new candidatesemerged: 2043, 2019, and 1970 BCE that are not listed here[14].
Most commentators commit themselves to the constellationof
Fang . They examine “automatically” a date in October.If one is willing to look at the putative course of eventswith some reservation, he comes across an eclipse less con-sidered: 15 September 1903 BCE (Figure 3).The hybrid solar eclipse flashed by in a distance of 20km from Anyi. The width of totality was as narrow as 15km. An observer would experience a full obscuration of thesolar disk for about 12 to 15 seconds, if he stood on thecentral line. The rotation rate of the Earth has been varyingto such an extent in the past that it must be taken into ac-count whether a solar eclipse could have been visible froma certain location or not.The Earth’s rotation is an exhaustive issue out of thescope of this paper — see the many works by Richard Steph-enson (e.g. [25]) or [12] for the most recent overview. Theclock error, termed ∆ T , is defined as the difference between the uniform timescale, based on the celestial motion of thesun and moon (Ephemeris Time, ET), and our standardisedcivil timescale which is used for the rotation of the earth(Universal Time, UT). For the remote past, that differenceis modelled by a parabola: ∆ T = ET − UT = − + ct with c ≈
32 sec/cy and t in centuries (cy) before 1820. Thisformula corresponds to a regular and systematic slow-down.It comprises the tidal friction as well as other seasonal ef-fects having influence on the behaviour of the earth’s ro-tation. We also know that there are irregular fluctuationssuperimposed on it and destroy a strict validity of the for-mula. If something unexpected happens, e.g. an earthquakeslightly altering the moment of inertia, a systematic errorenters and accumulates over time. It puts the backreckoninginto serious trouble: the path of the eclipse can be shifted toeither side of the mean longitude. An extrapolation beyond500 BCE turns out somewhat dangerous.Although a precise localisation of the eclipse track of1903 BCE cannot be given as long as the account in the Shujing lacks of a detailed description concerning totality,the social aftermath suggests a high magnitude of obscura-tion. Even without exactly determining the belt geographic-ally, the extrapolated ∆ T lies within the expected error bars.Fred Espenak’s map provide a ∆ T = ± ± ◦ in longitude.However, the observer does not need to find himself in-side the zone of totality to become deeply touched by thesudden loss of light thinking the end of the world is near.The ideal position will be even unlikely. An eclipse mag- . Khalisi (2020): The Solar Eclipse of the Xia Dynasty: A Review Figure 3: Central zones of two annular eclipses (yellow) and tree total eclipses (grey), one of which is of 1903 BCE. nitude of 0.996, as the average value provides, would dothe job, too. An incisive darkness, as in nighttime, in or-der to make stars visible, is not required. It rather supportsthe correctness of the empirical formula for extrapolationby Stephenson & Morrison [25].The advantage of this date is the reproduction of thehistorical corner stones. It would involve a liberate inter-pretation of the manuscripts. Historians will have to disen-gage themselves from the long counts of regents, and thedurations of their governances must be mapped in a morecompact way. Our new framework is outlined as follows:1953: Agglomeration of planets with a starting point of thecalendar — if this issue should matter at all;1925: then, a Great Flood with an unsuccessful emperorafter Yao (in accord with [32]);19**: then, an effective remedy of unknown duration underthe first Xia-Emperor Yu (see Table 1);1903: and, finally, the eclipse during the rulership of thefourth (?) Emperor Zhong Kang.Two pre-emperors and four Xia-emperors, at least,would have to be squeezed in a time interval of about 50years. Previously they were attributed much longer periodsof regency. So far, the time spans for the four Xia emper-ors are allocated at 45 years (Yu), 10 years (his son), 29(grandson), and 13 years (Zhong Kang himself), as given in the
Bamboo Annals . The German Wikipedia gives 58, 29,29, and 13 years, respectively, referring to ancient Chinesehistorians whose information will be outdated by now [30].The French version provides 8, 9, 29, and 13 years, respect-ively, crediting the historian Henri Cordier (1849–1925).More recently it was suggested that Emperor Yu’s regencymight have lasted between 1914 and 1907 BCE [16]. Thisabsolute timeframe would still conflict the flood and shouldbe shifted further back in time.That leads to the suspicion that the whole genealogy ofthe Xia should also be examined anew. Of the 17 knownnames (
Bamboo Books ), the shortest regency is 10 or 11years (king ≈
20 to 25 years betweenthe flood and Zhong Kang, four emperors could easily findtheir place.While the chronology of rulers is given back to the his-torians, astronomers have to accept some other weaknesses: . Khalisi (2020): The Solar Eclipse of the Xia Dynasty: A Review The constellation
Fang does not fully agree with the ac-count, and the literal minuteness must be abandoned in fa-vour of an approximate sighting. In September,
Fang is set-ting heliacally, i.e. one sees its final visibility at dusk beforedisappearing for the conjuction with the sun.This raises the general question about the scene of theeclipse: Would anyone really take care for celestial constel-lations while a dramatic phenomenon with a blackened sunfalls upon the unprepared and a civil commotion is goingon? Would the observer check for some stars at which the“non-harmonic meeting” of the sun and moon occurred? Oris it more likely that he guessed the time and place of theevent when the state of affairs have settled down? And whywould anyone keep such a detail for so many generationsafter him? — We believe that the dispensable note on theconstellation would rather be sought after the routine gradu-ally returned to everyday’s life [12]. It could have takensome weeks of regeneration when the description wouldbe put to record unless it did not slip at all. Then it wascopied without questioning by one writer after another tillit became an immovable part of the lore. At the restorationof the
Shujing , either the season or the constellation couldhave been readjusted with the calendrical precession of thewrong era.The cyclic day name does not match any dates, but itbecomes completely irrelevant because it was erroneouslyback-calculated in medieval times when the wrong positionof the sun in the sky was assumed. Besides that, for the an-cient people, any information about time had a much lowervalue than for us today. In the rural society, someone wouldnot hold on a certain day after, say, half a month or so. Heor she would just memorise the “dreadful event” and passit down in dramatic stories rather than cling to the trifles ofa constellation, or hours, or the celestial circumstances; es-pecially, when the storyteller is illiterate as the majority ofthe “common people” was. Such a loose perspective on thetexts leaves behind only the location of the observer as thesingle anchor for the determination of the eclipse. Unfortu-nately, even this is not explicitly specified, either.The solar eclipse of 15 September 1903 BCE is placedreasonably compatible with the planetary grouping and theGreat Flood. It produced the greatest magnitude in Anyibetween 2000 and 1850 BCE. In view of the badly transmit-ted information, it fulfils several logic criteria.Two more eclipses deserve a short notice (Figure 3).The track of totality on 3 July 1945 BCE (9:45 a.m. localtime) did cross the region, and its magnitude of 0.95 musthave left a noticeable effect in Anyi, provided that weatherpermitted. We disregard it because of the time-lag too shortsince the agglomeration of planets. The other suggestionby Kuniji Saito for 1912 BCE gives an annular eclipse withthe central path more than 600 km farther to the South (mag= 0.864). This option shows a somewhat better chronology,however, the obscuration happened during sunset at ≈ The most caveats were already touched in the preceding sec-tions. In general, accounts on eclipses in the Chinese chron-icles are both vague and sporadic before the 8th centuryBCE. In many cases the texts lack a correct understanding.In old times, people did not know how to describe such astrange phenomenon like the disappearance of the most im-portant celestial body [15]. Such incidents did cause muchpanic but they were not described in the prevalent terms.This makes it difficult to identify an eclipse in the records.Many events were surely observed without a message leftbehind, provided that weather permitted.Dates of any kind are scarce in the early history ofChina. Usually, the name of the regent was just recordedtogether with a keyword, as the example of He and Hoshows. Occasionally, a year appeared without the name ofthe regent. The earliest date that can be taken as a reliablepoint in the Chinese chronology is 841 BCE, the start of aninterregnum called “Gonghe Regency” [11]. Beyond thatdate we know almost nothing about the history except a fewnames of emperors and their relationships.The second unequivocal testimony about a solar ec-lipse goes back to 780 BCE. It shows up in the “Spring andAutumn Annals” (
Chunqiu ). The number of accounts oneclipses rises thereafter. So, there exists a gap of more thanone thousand years to the
Shujing eclipse. The lack of know-ledge must be ascribed to the burning of books in 213 BCE.According to some opinions, astronomical/astrologicalworks were not affected by this measure but rather those ofphilosophical and historical content [1]. Again others be-lieve that the huge gap is due to a discouragement inducedby the fate of He and Ho — scholars lost their motivationto show an active interest in astronomy, and this branch ofscientific occupation was avoided for a long time. On theother side, astronomy was of such great importance to theemperor that it guided his political engagements. He couldnot do without.Nonetheless, there is one subtlety in the eclipse accountof the
Shujing that should baffle: The phrase that the Sunand
Moon did meet. The Chinese would consequently haverealised that the moon was responsible for the eclipse phe-nomenon. Would this apply to those ancient times, too?Many other civilisations were far from such an awareness,for they thought that the moon would shine only at night. Itis our modern comprehension correlating it to an eclipse. Ifone tries to gain an imagination of the cosmic matters, it willbe necessary to observe several incidents of the same kindbefore the idea dawned that it was caused by the earth’scompanion. When the perception once entered belief, itwould be handed down to the successive events: the rela-tionship between the eclipse and the moon would be men-tioned more frequent. Hence, doubts are justified whetherthe text expresses the same what we would understand asan eclipse of the sun. If the so-called “meeting” was someother strange effect in the sky, verification fails. Therefore,the eclipse provides the only safe pillar for dating using as-tronomical means. . Khalisi (2020): The Solar Eclipse of the Xia Dynasty: A Review It becomes evident that there are manifold opinions onthe ominous eclipse in the
Shujing . They challenge the his-torical origin of the book itself. It seems more importantto examine the delivery of the report throughout the ages aswell as the survival of the book [6, 21]. In the 20th centurymore and more commentators pushed the eclipse of He andHo into the realm of legends. A lot of stuff therein appearsmystical. But, finally, a historical core accompanies everywriting.The experts on eclipse dating are remarkably silenttoday when it comes to a specification of a clear date.Though there is a permanent scientific improvement, e.g.in the physics of the earth’s rotation, and precision of com-putation, and methods of analysis, it all does not bail usout in this case as long as history does not bring forward adistinct range for the existence of the Xia dynasty. Most ofthe dates in Table 3 would yield an eclipse more than 200years before the appointed start of the Chinese calendar.The conflict affects any regent of that dynasty, too. If itshould ever be possible to date this
Shujing eclipse withoutdoubt, then one could immediately enlighten the completeadvent of China’s history.
We examined the sources on the oldest eclipse account inhistory and undertook efforts to question their authenticity.By considering other natural occurrences of mythologicalrank, we tried to date the legendary solar eclipse within thecontext of the Great Flood, the origin of the calendar, andthe first emperors of the first dynasty in China, the Xia.We came across the hybrid solar eclipse of 15 Septem-ber 1903 BCE. It must have swept through the area of Anyi,the supposed capital of the Xia. A correction of a fewminutes to the average ∆ T = Acknowledgements
The author thanks Jeremy Shears for important commentson improvement. Figure 1 is used under the licence CC-Lizenz BY-SA. The results are part of the Habilitation (ch.10.2) submitted to the University of Heidelberg, Germany,in February 2020 [12]. This article is now published “as is”on arXiv . The entire work was accomplished under severepressure.
References [1]
Brown, F. Crawford (1931) : “The eclipse in China”,
Popu-lar Astronomy 39 , Dec 1931, p567–573[2]
Chalmers, John (1861) : “Appendix on the Astronomy ofthe Ancient Chinese”, in:
Chinese Classics vol. III: TheShoo King , ed. by James Legge, Hongkong, 1861, p101–102; see [13][3]
Chambers, George F. (1909) : “The Story of Eclipses”, Mc-Clure Publ. Co., New York, 1909, p65–69[4]
Espenak, Fred (2006) : “Predictions for Solar andLunar Eclipses”, Eclipse maps courtesy of Fred Espenak,NASA/Goddard Space Flight Center, [5]
Fairbank, J.K. & Goldman M. (2006) : “China: Anew History”, 2nd ed., Harvard University Press, Cam-bridge/Massachusetts, 2006, p35; ISBN 978-0-67401828-0[6]
Fotheringham, John K. (1921) : “Historical Eclipses”,Clarendon Press, Oxford, 1921, p9–16[7]
Fotheringham, John K. (1933) : “The Story of Hi and Ho”,
Journal of the British Astronomical Association 43 , 1933,p248–257[8]
Gaubil, Antoine (1732) : “Traité de l’AstronomieChinoise”, in:
Observations mathématiques , ÉtienneSouciet, Paris, 1729–1732; cited by [29, 21][9]
Henriksson, Göran (2008) : “A new Attempt to date theXia, Shang and Western Zhou Dynasties by solar eclipses”,
Archeologia Baltica 10 , 2008, p105–109[10]
Herrmann, Dieter B. (1998) : “Die Jahrhundertfinsternis”,Paetec Verlag, Berlin, 1998, 1st ed., p22–23 (in German)[11]
Keenan, Douglas J. (2002) : “Astro-Historiographic chro-nologies of early China are unfounded”,
East Asian History23 , 2002, p61–68[12]
Khalisi, Emil (2020) : “Das Buch der legendären Finster-nisse”, Habilitation submitted to the University of Heidel-berg, 2020, ch. 10.2 (in German)[13]
Legge, James (1865) : “Chinese Classis”, vol. III (The ShooKing), Trübner & Co., London, 1865[14]
Liu, Ci-yuan (2002) : “Astronomy in the Xia-Shang-ZhouChronology Project”,
Journal of Astronomical History andHeritage 5 , No. 1/2002, p1–8[15]
Liu C., Liu X. & Ma L. (2003) : “Examination of earlyChinese records of solar eclipses”,
Journal of AstronomicalHistory and Heritage 6 , No. 1/2003, p53–63 . Khalisi (2020): The Solar Eclipse of the Xia Dynasty: A Review [16] Nivison D.S. & Pang K.D. (1990) : “Astronomical Evid-ence for the Bamboo Annals’ Chronicle of Early Xia”,
EarlyChina 15 , 1990, p87–95[17]
Oppolzer, Theodor Ritter von (1880) : “Über die Sonnen-finsternis des Schu-King”,
Monatsbericht der Königl. Preuß.Akademie d.W. , 1880, p166–185 (in German)[18]
Pang, Kevin D. (1987) : “Extraordinary floods in earlyChinese history and their absolute dates”,
Journal of Hydro-logy 96 , 15 Dec 1987, p139–155;doi: 10.1016/0022-1694(87)90149-1[19]
Pang K.D. & Bangert J.A. (1993) : “The Holy Grail ofChinese Astronomy: . . . ”,
Bulletin of the American Astro-nomical Society 25 , 1993, p922[20]
Pankenier, David W. (1981) : “Astronomical Dates inShang and Western Zhou”,
Early China 7 , 1981, p2–37[21]
Rothman, Richard W. (1840) : “On an Ancient Solar Ec-lipse observed in China”,
Memoirs of the Royal Astronom-ical Society 11 , 1840, p47–50[22]
Russell, Samuel M. (1895) : “Some Astronomical Recordsin ancient Chinese Books”,
The Observatory 18 , 1895 (No.231 + 232 + 234), p323–325 + p355–358 + p430–432[23]
Schlegel G. & Kühnert F. (1889) : “Die Schu-King-Finsternis”, Königliche Akademie der Wissenschaften, Ver-lag Johannes Müller, Amsterdam, 1889 (in German)[24]
Shaughnessy, Edward L. (2009) : “Chronologies of An-cient China: A Critique of the Xia-Shang-Zhou ChronologyProject”, in:
Windows on the Chinese World , ed. by ClaraHo, Lexington Books, Lanham/Maryland, 2009, p15–28;ISBN 978-0-7391-2769-8[25]
Stephenson, F.R. & Morrison L.V. (2005) : “Historical Ec-lipses”,
ASP Conference Series 335 , 2005, p159–180[26]
Stockwell, John N. (1895) : “Eclipses and Chronology”,
The Observatory 18 , No. 226/1895, p161–163[27]
Wang P.K. & Siscoe G.L. (1980) : “Ancient ChineseObservations of Physical Phenomena Attending SolarEclipses”,
Solar Physics 66 , 1980, p187–193; doi:10.1007/BF00150528[28]
Weitzel, Robert B. (1945) : “Clusters of five planets”,
Pop-ular Astronomy 53 , 1945, p159–161[29]
Whiston, William (1734) : “Six dissertations”, ed. by JohnWhiston, London, 1734, p195–209[30]
Wikipedia (2020) : https://en.wikipedia.org/ (Eng-lish + German sites), visited in February 2020[31] Williams, John (1863) : “On an Eclipse of the Sun recor-ded in the Chinese Annals . . . ”,
Monthly Notices of theRoyal Astronomical Society 23 , 06/1863, p238–242; doi:10.1093/mnras/23.8.238[32]
Wu Q., Zhao Z., Liu L. + 13 co-authors (2016) : “Outburstflood at 1920 BCE supports historicity of . . . ”,
Science 353 ,5 Aug 2016, p579–582; doi: 10.1126/science.aaf0842[33]
Wu Q., Zhao Z., Lui L. + 13 co-authors (2017) : “Re-sponse to Comments on ›Outburst flood at 1920 BCE . . . ‹”,
Science 355 , 31 Mrc 2017, p1382e;doi: 10.1126/science.aal1325 [34]
Zhang Y., Huang C.C., Schulmeister J. + 10 co-authors(2019) : “Formation and evolution of the Holocene massivelandslide-dammed lakes . . . ”,
Quaternary Science Reviews218 , 2019, p267–280; doi: 10.1016/j.quascirev.2019.06.011, 2019, p267–280; doi: 10.1016/j.quascirev.2019.06.011