Ancient Greek observations of T Coronae Borealis? Data and methodologies

Abstract

The “new star” which, according to Pliny the Elder, inspired Hipparchus to compile his star catalogue was not a comet and may have been the recurrent nova T Coronae Borealis. I first scrutinize Pliny’s sometimes over-interpreted report (surely based on a lost work by Hipparchus), then the evidence for Hipparchus’s dates. Data from Chinese annals, revised recently, point to a nova in the Libra–Scorpius region in 134 bc, probably too late in Hipparchus’s life to support Pliny’s claim. Likewise, the comet reported by Seneca and Justin, sometimes linked to this story, is now redated to 135 bc. After recalculating the interval between the 1866 and 1946 eruptions of T CrB, I confirm those posited for 1787 and 1217, using an English chronicle to refine the date of the latter. We know Hipparchus observed Corona Borealis; but the variation in intervals since ad 1217 rules out simple arithmetical reckoning of when T erupted during the 2nd century bc. More positive evidence comes from a suggestion that the naming of Coma Berenices in about 245 bc was inspired by an event in CrB. Given T’s roughly 80-year cycle, this suggests an eruption in the 160s, early in Hipparchus’s active life.

Dedicated to the memory of my teachers at the Royal Grammar School, Newcastle upon Tyne (founded 1525)

Introduction

Public and media interest in T Coronae Borealis (T CrB), the long-period dependent binary star and recurrent nova—briefly visible to the naked eye in 1866 and 1946—has been stimulated in the last few years by the prospect of its third modern eruption, expected during 2025–2026. In an earlier note, ¹ I raised the possibility—hitherto unconsidered—that the “new star” whose sighting, according to Pliny the Elder (mid-1st century ad), prompted the Hellenistic astronomer Hipparchus of Nicaea (mid- to late 2nd century bc) ² to embark upon his great star catalogue, was none other than T CrB. In this fuller study, I aim to build the strongest case possible, while recognizing that certainty is unattainable.

The paper first scrutinizes Pliny’s words to establish new limits of interpretation. After briefly assessing the chronological evidence for Hipparchus, it examines the data sometimes adduced in support of the view that the event underlying “Hipparchus’s star” appears in medieval Chinese annals and a late Latin compendium of omens. I then re-examine the eruptions of T CrB in 1946 and 1866, with the aim of calculating the interval between them more precisely. The investigation then turns to the credible evidence for eruptions in 1787 and 1217; for the latter, I resurrect a neglected evidential link that allows us to suggest a more precise date for the event.

I return to ancient times to review briefly what we know of Hipparchus’s relevant observations from his surviving and fragmentary works; and to offer a detailed analysis, based on the intervals between the medieval and modern eruptions, demonstrating that no method of arithmetical retrojection back to the 2nd century bc can reliably show whether Hipparchus is likely to have had an opportunity to see T CrB. Finally, however, and more positively, I find unexpected value in hints from Hellenistic and Roman Republican poetry that T may have erupted in the 240s bc, making a subsequent eruption during Hipparchus’s early career a distinct possibility.

The inquiry raises issues of how to combine scientific data with historical source criticism beyond factual “documents.” Human-centred, culturally sensitive interpretation—not merely a positivistic analysis of recorded evidence—is an indispensable adjunct to astrophysical data. Furthermore, astronomers and historians need to be extremely wary of modern translations of medieval and ancient sources.

The evidence of Pliny the Elder

The Roman author Pliny the Elder, who died in the eruption of Vesuvius in ad 79, writes that Hipparchus of Nicaea, the famous astronomer and geographer of the second half of the 2nd century bc, saw a “new star” (novam stellam; Natural History 2. xxiv. 95); and that this prompted Hipparchus to wonder whether other unfixed stars existed (besides the planets), spurring him on to compile his famous star catalogue which was the foundation of the great Greek work of the Roman period by Ptolemy of Alexandria.

The wording of Pliny’s text varies slightly between different surviving medieval manuscripts, and also between different modern editions, though the words novam stellam are common to all. Here is the key sentence, with a possible translation: ³

novam stellam et aliam in aevo suo genitam deprehendit; eiusque motu, qua fulsit, ad dubitationem est adductus anne hoc saepius fieret, moverenturque et eae quas putamus adfixas.

He discerned a new and different star born in his own time; and was led, by its movement in the place where (or “by the manner in which”) it shone out, to wonder whether this might not occur quite often, and whether even those stars we think are fixed might not be moving. ⁴

The passage admits of different interpretations, the chief issue being whether Pliny thinks he is describing a star (in our sense) or a comet. Greek and Roman writers’ paraphrases of their sources are often far from exact, especially when a change of language is involved. For all his encyclopaedic knowledge, Pliny does not observe modern standards of scholarly attribution, not even telling us where and in which of Hipparchus’s works (all but one of which are now “fragmentary”—that is, surviving only through later quotation, paraphrase, or citation) the discovery was mentioned. Until recently, both classicists and historians of astronomy have tended to castigate Pliny for his errors and his failure to give precise source citations, without grasping that he was more a literary artist than a scientist. ⁵ In fact, by the standards of his age he is unusually conscientious: in the first book of Natural History he lists every author he has consulted for each of the remaining 36 books. For book 2, he lists Hipparchus first among his sources, and there is no reason to doubt that he consulted a work or works by the earlier writer directly, not through an intermediary’s adaptation or report—though he may have cited or paraphrased particular passages from memory.

As a working principle, we should assume that the reason why Pliny says Hipparchus saw a new star is that he has read it in Hipparchus’s own words. Unfortunately, as already noted, only one of Hipparchus’s writings survives complete: his Commentary on the astronomical writings of Eudoxus of Cnidus (4th century bc) and Aratus of Soli (early 3rd century bc). This Commentary does not refer to any “new star” in Corona Borealis. ⁶

Star or comet?

After stellam, “star,” manuscripts of Pliny include the words et aliam (“and (an)other”), resulting in the literal sense “a new and other star” or “a new and different star.” Some editors remove these words, regarding them as intrusive (that is, deriving from an interlinear or marginal comment or addition made by the scribe of an earlier copy of the text); and one scribe thought that vel aliam, “or (an)other,” was a preferable reading. ⁷ In the preceding lines, Pliny has been talking about Hipparchus’s cosmology, not about the discovery of stars or other heavenly bodies, so “other” and “another” seem inappropriate. “Different,” rather, should be the meaning. Different from what? In the preceding paragraphs, he has been talking about comets, though not about discoveries of comets as such, so the point of aliam seems to be a contrast with comets.

J.K. Fotheringham, evidently using an edition of Pliny containing et aliam, offers “yet another new star” (italics mine); ⁸ but this strains the Latin. He presumably takes the category of stars here to include comets, since the thrust of his paper is that a great comet documented (as he believed) in Chinese records for 134 bc (year –133) was Hipparchus’s new “star” (for further discussion of these records, see below). D.R. Dicks, too, ⁹ on balance marginally favours a comet over a nova, though the latter is favoured by Jean Beaujeu, who prints et aliam and translates as découvrit une étoile nouvelle, différente des comètes (“discovered a new star, different from the comets”). ¹⁰ Otto Neugebauer doubts that Beaujeu is right to translate et aliam in this way; ¹¹ but it is hard to see what else its import could be, comets having been the subject of Pliny’s preceding pages. ¹²

We do not need the words et aliam or vel aliam to tell us that Pliny, at this point in his text, puts aside comets and begins discussing Hipparchus and stars (in the narrow sense). The rest of the immediate context makes this clear, so much so that Beaujeu is rightly confident in stating that l’étoile nouvelle découverte par Hipparque n’était pas une comète (“the new star discovered by Hipparchus was not a comet”). ¹³ (Yet a recent, standard compendium of historical comets misleadingly includes the Pliny passage among the primary sources for the great comet hitherto dated to 134 bc.) ¹⁴

We should assume, then, that Hipparchus’s star was some heavenly body other than a comet. Gerhard Winkler and Roderich König cite Dicks’s Text I ¹⁵ for Hipparchus’s star being a fixed star in Scorpius; ¹⁶ but this source is the same Pliny passage, which names no constellation or any region of the celestial sphere. Dicks is presumably making an inference from a Chinese record (below) referring to a region that includes Scorpius.

A few lines below the passage quoted, Pliny goes on to explain the purpose of Hipparchus’s catalogue as ut facile discerni posset ex eo non modo an obirent ac nascerentur, sed an omnino aliquae transirent moverenturque, item ac crescerent minuerenturque (“in order that with its help it could easily be discovered not only whether they (the stars) died and were born, but even whether some changed their position and moved, as well as grew and diminished”). This, too, seems more easily reconcilable with stars than with comets. At first sight, it could cover both the precession of the equinoxes (now called “axial progression”), which Hipparchus himself discovered, and the phenomenon of variable stars. Unless, however, Pliny’s report of the nova stella is confused, it is not about precession. A nova would be of no assistance in detecting this: in the pre-telescopic age, observations separated by decades were necessary.

“Movement”

The word motu, “movement,” in Pliny’s text has caused some difficulty. Since he has stopped talking about comets, the word must refer to the behaviour of some other kind of body. One might think of the slow drift of a bright asteroid; or perhaps he has transmitted Hipparchus’s report of a change in brightness rather than position. In his translation, Beaujeu gives constatant que le point où elle brillait se déplaçait (“noticing that the point where it was shining was changing its position”), ¹⁷ glossing this in his commentary with l’expression qua fulsit semble désigner un déplacement dans l’axe de visée, c’est-à-dire la traduction cinétique d’une variation d’intensité (“the expression qua fulsit seems to refer to a displacement in the axis of sight, that is to say the kinetic translation of a variation of intensity”); ¹⁸ an alteration of magnitude in our sense of a star’s brightness—or “size,” magnitudo, as Pliny would have called it—perhaps being understood by the observer, reasonably enough, as a change in the angular space occupied by the star (and perhaps in its distance from us). ¹⁹ Neugebauer notes that Pliny does use verbs of movement for changes in fixed stars. ²⁰ It is difficult, moreover, to see why Hipparchus would have posed his problem as extending to fixed stars if the object he saw was manifestly changing position.

After qua, which in this context must, by itself, mean “in the place where” or “by the manner in which,” ²¹ annotators of several manuscripts have added diē (“day”), which changes the meaning to “on which day” or “on the day on which.” Most editors, though not Winkler and König, reject this. It is probably a scribal “gloss” designed to explain something where an earlier copyist was not satisfied that the sense was complete; presumably it was intended to convey the instantaneity of the star’s appearance.

Sudden, momentary, or sustained?

Whether we admit diē or not, however, is of limited importance, for there is independent reason to believe that the star brightened suddenly. Judging by the verb that Pliny uses to describe this event, it seems to have been instantaneous rather than gradual: fulsit (perfect tense), “shone out,” rather than e.g. fulgebat (imperfect tense), which would convey “was shining,” “began to shine,” or “used to shine,” expressing an event of some duration. This does not sound like a comet, a phenomenon with which the Greeks and Romans were perfectly familiar and which would not have held any surprise value for Hipparchus, as the story requires.

Hipparchus’s dates

To discover the likelihood that Hipparchus’s “new star” was T CrB, we first need to know when Hipparchus was active. Dicks argues convincingly, following Fotheringham, that Hipparchus’s observations included those of spring and autumn equinoxes back to 162 bc, a case he supports by detailed examination of sources. ²² Among the more important of these is the testimony of Theon of Alexandria (4th century ad), who states that these autumn and spring equinoxes “had been observed accurately by Hipparchus”; and Dicks quotes Fotheringham’s wry comment that they must be Hipparchus’s own, for “Of which of his predecessors would so severe a critic as Hipparchus have said that his observations were made with the greatest possible accuracy?” ²³

Neugebauer takes a narrower view—dating to 146 bc (–145) the earliest observation in Ptolemy’s Syntaxis that seems definitely to derive from Hipparchus, with others dated 134 bc and 127/6 bc ²⁴ —but does not appear to have considered the whole evidential context. Dicks also cites evidence that Hipparchus saw an eclipse in 141 bc (year –140). ²⁵ Among other observations, René Bourtembourg has argued persuasively, on the basis of Ptolemy’s account of a quadrilateral of stars within Virgo (Almagest, 7. 5), that Hipparchus saw Uranus there in 128 bc (year –127). ²⁶ (Bourtembourg does not suggest that Virgo “17,” as Ptolemy calls it, was Hipparchus’s “new star”; it would be hard to square with Pliny’s fulsit, given that the planet’s magnitude was only 5·4 at the time and barely changing. As in the discussion of “movement” (above), if Uranus were the “new star,” why would Hipparchus, as reported by Pliny, have posed his research question as being about fixed stars, when Uranus moves c.4½ degrees a year against the background of stars, albeit with standstills, one of which occurred around the date proposed by Bourtembourg?)

It seems safe to accept that Hipparchus is likely to have been an active observer at least from 162 to 127/6 bc, a period of some 35 years; a credible span for a scholar who wrote copiously. It is worth observing, as a principle of method, that modern publications often cite one of these contradictory dates or date ranges at second or third hand as if they were established facts, without presenting alternatives and giving no indication of what evidence lies behind them, or that they are modern constructions and contestable. Another such factoid is the date of 129/8 for when Hipparchus began compiling his catalogue, sometimes reported as definite but resting on questionable grounds. ²⁷

Given T’s periodicity of c.80 years, the probability that it erupted during Hipparchus’s active years is a priori roughly 40%–45%.

Chinese comets and Roman omens

Hipparchus’s “new star” has been linked to the report in Chinese annals of a comet with a tail spanning a large arc of the heavens. This hypothetical link has repeatedly appeared in scholarship as if it gave a firm date for when he began work on his catalogue, its epoch, or his life dates. Unfortunately, not only does the hypothesis contradict our reading of Pliny (above), which found that the “new star” was not a comet; it is also, inconveniently for the present investigation, impossible to connect with Corona Borealis. In fact, as we now know, the hypothesis depended on an accidental conflation of two separate events.

In the late 19th century, John Williams, fsa, established—in a self-funded publication with plates drawn by his own hand—that Chinese records mention a “strange star” seen in 134 bc (year –133) in the Stellar Division named Fang, a region of the heavens including groups of stars south of the celestial equator in the modern constellations of Libra, Lupus, Ophiuchus, and Scorpius. ²⁸ The term “strange star” might better be rendered as “extraordinary star,” but actually translates a Chinese term meaning literally “guest star” (rendered as étoile-hôte in French). ²⁹ Fotheringham, following Édouard Biot, states that the term normally denoted what was thought to be a comet without a tail; ³⁰ and, further, notes that comets appear in Chinese records for 134–129 bc (–133 to –128) and 122–117 bc (–121 to –116), equating the former with the “guest star.”

J.T. Ramsey, however, has now shown that the great comet should be backdated to 135 bc (–134) and identified with one mentioned by Seneca the Younger (1st century ad; Quaestiones naturales, 7. 15. 2) as well as with that reported later by Justin (date uncertain; later than Seneca) in his abridgement of the Philippic History by the Gaulish historian Pompeius Trogus (1st century bc; 37. 2. 1–3). Ramsey also shows that the “guest star” of 134 bc (–133) is more likely to have been a nova than a comet, ³¹ and suggests that this may be Hipparchus’s “new star.” This would exclude T CrB from that honour, for although Corona Borealis and Fang are both summer groups, having a similar Right Ascension (Corona Borealis c.16 hours, Fang 15–17 hours), the Fang constellations are in the southern hemisphere (Libra and Scorpius being zodiacal, with declinations around 10–40 °S) while CrB (as defined today) has a declination range of 26–39 °N. That said, the Chinese record dates the sighting to between 22 June and 21 July 134 (–133), and all these constellations are visible through summer evenings from the Mediterranean, Libra and Scorpius setting around midnight; ³² so if there was a nova in that region, Hipparchus may have seen it.

Fotheringham links the great Chinese comet of 134–129 bc—now re-dated to 135 bc (–134) by Ramsey—to a report in the collection of Prodigies by the Roman author Julius Obsequens (c. ad 400). Obsequens reports, again under the year we designate 134 bc (–133), that at Amiternum in east-central Italy sol noctu visus, eiusque lux aliquandiu fuit visa (“the sun was seen at night, and its light was seen for a considerable time,” chapter 27). ³³ Some view this as a comet, ³⁴ though Fotheringham concedes that, were it not for the coincidence of dates, it would be hard to say so; but Ramsey, as we have noted, places the comet in 135 bc (–134) and makes the event of 134 bc (–133) a probable nova. So we may link this to the event in the Chinese annals, and may speculate on the basis of Obsequens’s reference to “the sun” that this was a very bright star indeed. The year 134 bc is rather late in Hipparchus’s career for him to have begun and finished his star catalogue; but it is not impossible, and this event must remain a theoretical possibility for the “new star.”

Modern observations of T CrB

In investigating the past history of T Coronae Borealis, we may hope to achieve greater precision from revisiting modern records. ³⁵ It must be borne in mind that a daylight eruption will never be seen immediately in the absence of instrumentation of today’s power; nor will the first reported sighting of a new star necessarily coincide with, or even follow closely upon, its actual eruption. Visibility also requires Corona Borealis to be above the horizon, which it is for c.17 hours at the latitude of the southern British Isles, for c.14½ hours in Athens, and for c.14 hours at Alexandria. ³⁶

The 1946 eruption

The first observation of the most recent eruption appears to have been made on 8 February 1946 at Shimanovsky (now Shimanovsk) on the Trans-Siberian Railway in eastern Russia (52° 00′ N, 127° 40′ E) around 1900 UT (i.e. 0400 local time on 9 February) by A.S. Kamenchuk, who gave the star’s magnitude as 1·7. ³⁷ The Julian date is 2 431 860 29167, which we may round to 2 431 860 3 since the report does not include an exact time, and because we do not need greater accuracy for present purposes. There is no information that would allow us to estimate how long after eruption the observation took place.

Two British observers saw it later the same night, early on 9 February. The first was Michael Woodman, ³⁸ the second N.F.H. Knight. ³⁹ Together, the three observations suggest that T was first observed around its maximum, and that it was already declining within a matter of hours.

The 1866 eruption

Some eighty years earlier, the village of Tuam, in County Galway, Ireland, was the location of the first ever documented naked-eye sighting of T CrB. William Huggins, frs, reported to Astronomische Nachrichten that he had heard on 16 May 1866 from John Birmingham, who had seen the star there late on the evening of 12 May. ⁴⁰ Birmingham himself later wrote to the editor of the Nachrichten on 7 July saying that he had seen it first between 2330 and 2345 “Tuam time.” ⁴¹ As Tuam is 8° 51′ W of Greenwich, local time (assuming it was local mean time) will have been c.35 minutes behind GMT, so that 2330–2345 on 12 May equates to c.0005–0020 GMT on 13 May.

In a report published in the Nachrichten a few weeks later, the renowned and indefatigable director of the Observatory at Athens, J.F. Julius Schmidt, reported that T CrB had not been visible (sc. to the naked eye) at 2145 Athens mean time (“mittl. Zeit Athen”) on 12 May. ⁴² Athens lying 23° 44′ E of Greenwich and thus c.1 hour 35 minutes ahead, this equates to 2010 GMT, some four hours before Birmingham’s observation. We may therefore estimate the time of eruption as 2210 ± 2 hours GMT on 12 May 1866, and the Julian date as between 2 402 734 34028 and 2 402 734 50694, with a midpoint of 2 402 734 42361, which we may round to 2 402 734 4.

Schmidt then observed the star on the evening of the same day, 13 May, and found that it was already dimming. ⁴³ He followed the star for several weeks, and drew a light curve showing that it had dimmed to magnitude 6 in seven days. ⁴⁴

1866 and 1946 compared

On the basis of these two modern events, and of a number of other observations after both maxima, E. Pettit found that the two series of light curve data matched closely. In each case the star remained visible, in principle, to the naked eye for around a week. ⁴⁵

We now understand that T CrB is a binary occluding variable, a pair of stars comprising a red giant and a hot white dwarf. Normally it is around magnitudes 9 to 10, but the white dwarf is continuously drawing material out of the red giant to form an accretion disc. Every 80 years or so, the system becomes unstable, the white dwarf swells up massively, and the star brightens to magnitude 2 or 3 for a few hours.

The interval between the two modern eruptions was 29,125·9 days (79 years c.271·8 days ~ 79·74 years). ⁴⁶ Schneider observes that this is almost exactly 128 orbital periods of the star at 227·53 ± 0·02 days per orbit. ⁴⁷ That orbital period itself is an average, though it varies by only a few hours. ⁴⁸ Schneider assumes that the interval between eruptions will be a whole number of orbital periods; but the basis for this assumption seems doubtful. Why should the accretion disc reach a critical disequilibrium after an exact number of rotations—especially as the white dwarf’s orbit is almost perfectly circular? The nearly integral number is presumably a coincidence. ⁴⁹ In attempting to identify or reconstruct outbursts of T before 1866, we will do better to be guided by the absolute intervals between observed and reconstructed eruptions.

Reconstructed observations

The 1787 eruption

Bradley Schaefer has assembled convincing arguments that an eruption of T CrB was observed some days after maximum brightness by the Revd Francis Wollaston, frs (1731–1815); ⁵⁰ presumably at Chislehurst, Kent, where he was Rector from 1769 until his death. ⁵¹ He made his observation on 28 December 1787. Since the limiting magnitude of his equipment was 7·8, T must have been in eruption and the maximum must have occurred about 8 days earlier, around 20 December (Julian date at 00 hours UT 2 374 101 5). This would put the maximum c.28,632·9 days (78 years c.147·0 days ~ 78·40 years) before the 1866 eruption—a shorter interval than that between the 1866 and 1946 outbursts by 1·34 years.

If the deduced observation is accepted, it shows that T CrB does not follow an exactly regular cycle.

The 1217 eruption

In Sky & Telescope for 1967, in a letter to the editor—one of many that she wrote to magazines and journals during half a century—the well-known and highly scholarly amateur astronomer Cicely Botley from Tunbridge Wells, Kent, ⁵² notes a late medieval report of a variable star or nova, which she suggests was T CrB. ⁵³

Botley’s letter, and the 1881 paper in German that she had read, ⁵⁴ quote from a Latin chronicle written in 1229–1230 by Burchard, provost of the Abbey of Ursberg in Bavaria. He reports a celestial event he had witnessed in “Ariadne’s Crown” (the old name for Corona Borealis) in 1217. Schaefer, like Botley, and citing a suggestion by R. Webbink, identifies this as an eruption of T CrB (though by an oversight he does not mention Botley’s suggestion). ⁵⁵ I quote the text with my own punctuation and a new translation. It stands at the end of a passage introduced with Anno Domini millesimo et xvii, “in the year ad 1217”:

Eodem anno tempore autumpnali, hora vespertina post occasum solis, in quadam stella in occidente visum est signum mirabile. Nam—cum stella illa, posita versus austrum, aliquantulum declinans in occidentem, in directo sideris illius quod vocant astrologi coronam Ariadnae, sicut nos ipsi annotavimus, antea erat parva et post ad parvitatem rediit sed tunc maiori lumine refulsit, visusque est ab ea ascendere versus altitudinem firmamenti quidam radius valde clarus, quasi trabes magna et alta, et hoc per multos dies, ut predictum est, tempore autumpnali sero visum est—post paulatim defecit et ad suam parvitatem stella rediit. Predicatores quoque his temporibus multa alia asserebant contigisse signa in celo et terra, quae longum esset enumerare et huic brevitati annectere. ⁵⁶

And in the same year, in the autumn time, at the evening hour after sunset, in a certain star (stella) in the west a marvellous sign was seen. For—although (?) ⁵⁷ that star, situated towards the south but diverging very slightly towards the west, in the line of (in directo) that constellation which astronomers call Ariadne’s Crown, as we ourselves recorded, was previously small and afterwards returned to smallness but at that time shone out again with a greater light, and (?) there was seen to arise from it into the heights of heaven a certain ray, shining brightly like a great high beam, ⁵⁸ and this, as mentioned earlier, was seen for many days, late in the autumn time—afterwards the star gradually declined and returned to its smallness. Preachers also declared at these times that many other signs had occurred in Heaven and Earth, which would take a long time to list and include in this short account.

Burchard’s use of the verb refulsit, “shone out again,” recalls Pliny’s fulsit: a distinct, instantaneous occurrence. In other respects, however, there are a few warning flags for the sensitive linguist. (1) Burchard says the star was “previously small” (parva), meaning faint; yet on the basis of the two modern sets of observations it should have been quite invisible to the naked eye. (2) Taking hoc, “this,” to refer to all the phenomena previously mentioned, ⁵⁹ the text asserts that the star and the associated “beam” were visible “for many days” (per multos dies), which is difficult to reconcile with naked-eye visibility of around a week, as in the two modern eruptions of T. (3) Burchard says the star “returned to its smallness” (ad suam parvitatem rediit); but it ought to have disappeared entirely.

All three questionable remarks, however, can be explained from a source-critical point of view in a variety of ways. Perhaps the most plausible reason for Burchard’s inaccuracy would be that, when writing a dozen years after the event—though apparently having made notes at the time ⁶⁰ —he suffered from faulty memory or elaborated, consciously or not, what he had seen. He may also have been thinking, to some degree, of the kind of report he was used to reading and that he imagined his readers might expect to be told.

As to the “ray” or “beam,” Schaefer opines that such reports are unreliable; even if a report reflects a real phenomenon, it could be one of a number of kinds of event. But we can be more optimistic than that. Schaefer might have mentioned, as Botley does, that by an extraordinary coincidence an English chronicle of the very same year, 1217, mentions a celestial event in the same autumnal season.

The Annals of Dunstable Priory record that in vigilia Simonis et Judae, visum est cuidam canonico de Dunstaplia videre in aere quandam crucem immensam transire cum ingenti gloria ab occidente versus partes orientales (“on the eve of (the feast of) Simon and Jude, it seemed to a certain canon of Dunstable that he saw a certain vast cross in the air pass with great glory from the west to the eastern parts”). This gives a definite date of 27 October. ⁶¹ Whether this was, as Botley suggests, “the zodiacal light, somewhat enhanced by auroral activity” or some other phenomenon, the similarity to Burchard’s report of a “ray or beam,” seen in the west (since it was linked to Corona Borealis on an autumn evening), seems too remarkable to be mere coincidence, even if it does not fix an exact date for the eruption of T Coronae Borealis. Evidently no eruption of T CrB was noticed from Bedfordshire. A series of auroral displays on successive nights would be compatible with Burchard’s “many days”; perhaps visible from Dunstable on only one night because of weather conditions. If there was an eruption, its maximum may have occurred, if not on 27 October itself, up to a few days before or after this date.

The most persuasive feature of Burchard’s narrative is the specific mention of Ariadne’s Crown, Corona Borealis. Additionally, one is encouraged towards belief by the rather precise report of the time of night when the observations were made. Given that Ursberg is at latitude 48° 16′ N, Corona Borealis on and around 27 October should have been visible from Bavaria for around 4 hours after sunset before setting in the north-west. ⁶²

Taking 2200 hours local time on 27 October as a notional time would place the event at Julian date 2 165 860 41667, which we may round to 2 165 860 4; that is, 208,241 days (570·15 years) before the probable date of the 1787 eruption, equating to seven eruption cycles at an average of 29,749 days (81 years 164 days ~ 81·45 years). ⁶³ This is 2·05 years longer than the 1787–1866 interval, and 1·71 years longer than the 1866–1946 interval. Such variation is not surprising; stars of this kind may vary not only in brightness but also in intervals between maxima. ⁶⁴

It appears, then, that the interval between eruptions was longer, on average, during 1217–1787, shorter from 1787 to 1866, and slightly longer again from 1866 to 1946. It may be that between 1217 and 1946 the star’s periodicity was reducing, albeit irregularly. A possible implication is that in the centuries before 1217 its average periodicity was longer still, which of course affects any attempt to “predict” an eruption that might coincide with the working life of Hipparchus. Alternatively, the periodicity between the 2nd century bc and the 12th century ad was stable, or increasing, or irregular.

Which brings us back to the Hellenistic period.

Hipparchus’s observations of Corona Borealis

As it happens, we know Hipparchus observed Corona Borealis, or Stephanos (“The Crown”) as he called it, in detail. The evidence is found in several passages of his Commentary on Aratus and Eudoxus. While in some he only reports information from those two authors, in others he reports his own observations, as here:

As the Crown rises, there rises with it the section of the zodiac from the 27th degree of Virgo to the middle of the 5th degree of the Claws (Scorpius). [. . .] And, of the Crown, the first star to rise is the one preceding the brightest, ⁶⁵ while the last is the most northerly of those lying east and north ⁶⁶ of the brightest. [. . .] And the Crown rises in two parts of an hour (i.e. two-thirds).

(Commentary, 2. 5. 2, my trans.) ⁶⁷

In this and other passages, Hipparchus refers to the stars we call α, β, γ, δ, ε, and ι CrB, giving detailed positions as well as the sequence in which they rise (as above) and set. ⁶⁸

Several of the stars reappear in a passage of Greek discovered recently, using modern technology, beneath a later Syriac text in a palimpsest (reused manuscript). Its editors argue that the words are part of the entry on Corona Borealis from Hipparchus’s star catalogue, ⁶⁹ where entries take a different form from those in the Commentary. The association with Hipparchus has been questioned, ⁷⁰ but at least two other specialists promptly accepted the identification, ⁷¹ and the editors have now published a firm rebuttal of their critics. ⁷² The passage gives the positions of α, ι, δ, and probably π CrB. I give a slightly simplified translation (italicized letters and words in parentheses are mine):

Corona Borealis, lying in the northern hemisphere, in length spans 9¼° from the first degree of Scorpius to 10¼° in the same zodiacal sign. In breadth it spans 6¾°, from 49° from the North Pole to 55¾°. Within it, the star to the West (i.e. β CrB) next to the bright one (α) leads (i.e. is the first to rise), being at Scorpius ½°. The fourth star (ι) to the East of the bright one is the last (to rise) [— text damaged —]10 49° from the North Pole. Southernmost (δ) is the third counting from the bright one towards the East, which is 55¾° from the North Pole.

The rediscovered text does not mention a star we can identify as T, but it would have been surprising if Hipparchus listed a star that had once been visible briefly but then dropped below naked-eye visibility. The precision and richness of the information in both his writings does, at least, suggest that if T CrB had been there to be seen during his working career, he could well have seen it. (That is not to exclude the probable nova of 134 bc in the Libra–Scorpius region, discussed earlier.)

When did T CrB erupt in the 2nd century bc?

To establish the likelihood that Hipparchus could have seen T CrB, the simplest response—before Schaefer’s discovery of eruptions in 1787 and 1217—would have been to take the only securely timed interval, the most recent, and suppose that eruptions before 1866 took place at an interval of 29,125·9 days. On this basis, 26 cycles back takes us to c.5 October 129 bc (year –128) within Hipparchus’s lifetime, but rather late within it if we credit the story that sighting the “new star” prompted him to compile his catalogue.

The addition of an eruption around 20 December 1787 creates a slightly shorter inter-eruption interval of 28,632·9 days. We could then take the average of the two latest intervals, 28,879·4 days. Retrojecting 24 and 25 cycles of this length from 1787 takes us to 23 April 111 bc (–110) and 30 March 190 bc (–189) respectively—neither of them a good fit for Hipparchus’s working life.

Then again, if the 1217 eruption is accepted, a Julian date around 27 October, as noted above, dictates that the average of the next seven cycles was greater than the eighth and ninth cycles (1787–1866–1946). We then have several options.

• (a) We may take the average of all the intervals from 27 October 1217 to 13 May 1946, which is c.29,555·54 days (80 years 336 days ~ 80·92 years). Reckoning back as before, 17 cycles “predict” an eruption c.15 November 160 bc (year –159), which would fall at the beginning of Hipparchus’s career and potentially validate Pliny’s story.

• (b) If the average of c.29,748.73 days (81 years 164 days ~ 81·45 years) per eruption between 1217 and 1787 held good in the centuries before 1217, one predicted date in the late Hellenistic period fits Hipparchus’s chronology well, namely c.9 March 168 bc (–167), again near the likely beginning of his career. But there is no reason to suppose that the average inter-eruption period before 1217 was the same as in the seven cycles after 1217. Caution is necessary: even a modest variation above or below 81·45 years in the average length of cycle will change the “landing-point” by years or decades.

• (c) A more sophisticated option would be to assume that the periodicity of T CrB has been decreasing overall since 1217—not consistently, however, since the most recent interval (1866–1946) was longer than the one before it (if we accept the 1787 event). We can then project the trend back in time before 1217 until we reach Hipparchus’s era. In a linear trend line, ⁷³ the retrodicted period reaches c.86 years by the 2nd century bc and yields a theoretical eruption date in 127 bc (–126). Not only is this late in Hipparchus’s career, it has little if any evidential value, given the uncertainties surrounding all the data. As under (b), slight changes in the trend will yield widely varying dates for events 13–14 centuries before 1217.

Indeed, assuming Hipparchus’s active career lasted about half a T CrB cycle, there is a roughly 50% chance that any figure we use for the eruption cycle will land somewhere in his adult life! We are not entitled to choose the trend line to fit the incomplete historical record.

A final point. There were quite likely other naked-eye novae in Hipparchus’s era that could be his “new star” but of which we know nothing. Indeed, we have already seen plausible evidence of a nova in 134 bc (year –133) which, though relatively low in the summer evening sky, may have been visible to an observer at Rhodes or Alexandria. More generally, seven occurred in, for example, the 20th century, most of them with longer periods of visibility than T CrB. ⁷⁴

To conclude this section: the uncertainty surrounding the interval between eruptions, and indeed the proven variation in the interval, mean that we cannot predict with any accuracy when an eruption or eruptions of T CrB took place 21 or 22 centuries before the present. ⁷⁵ We must content ourselves with acknowledging the possibility indicated by the historical evidence.

There remains, however, one angle to consider, which may allow us to build a more plausible case that Hipparchus’s star was T Coronae Borealis.

Conon, Coma Berenices, Callimachus, Catullus, and Corona Borealis

Corona Borealis was not the only star group to undergo alteration in the early to middle Hellenistic period (3rd and 2nd centuries bc). The other change, however, was manmade.

At Alexandria in 245 bc or shortly after, the astronomer Conon of Samos identified a group of stars that rises after Leo and before Boötes—and thus not long before Corona Borealis—as the catasterism (celestial elevation) of a living queen’s lock of hair (κόμη–komē); today we call it Coma Berenices (Com). ⁷⁶ In this final section, I explore the possibility that an eruption of T Coronae Borealis was somehow connected with the naming of this “new” constellation.

The queen was Berenice II; her husband was Ptolemy III, the third Macedonian king of Egypt from 246 bc. Like many ancient Macedonian rulers, he sought to consolidate his kingship with military success at the outset of his reign, campaigning in western Asia against the rival Macedonian dynasty of the Seleukid kings, in what we call the Third Syrian War. A series of events surrounding his return from that war were immediately mythologized by the renowned court poet Callimachus (c.310–c.240 bc), who made it the final poem of his four-book collection in elegiac couplets known as Aitia (Explanations, of mythical character). His poem, in turn, was translated into elegant Latin elegiacs by the late Republican poet Catullus (c.84 to c.54 bc), and is now number 66 in his surviving oeuvre.

Unfortunately, both poems survive in a damaged state: that by Callimachus on a fragment of papyrus from Roman Egypt; that by Catullus in his scanty manuscript tradition. Combining what they tell us, we learn that the queen vowed to cut off her hair (or some of it), and dedicate it in the temple of Arsinoë Zephyritis, ⁷⁷ if the king returned safely. He did so, in 245; she did dedicate the hair; and it vanished from the temple. Thereupon the court astronomer, the aforementioned Conon, identified a group of stars—previously regarded as part of Leo, perhaps the tuft at the end of his tail—as the hair which the West Wind (Zephyros) had purloined and presented to Aphrodite, the goddess of love, whereupon she had placed it in the sky. The star group was now named komē Berenikēs, Berenike’s Tresses.

With its brightest star (α Com) at magnitude 4, slightly fainter than those of the Hyades and the brightest of the Pleiades, the constellation is often thought unimpressive. Graphic artists often rise to that undemanding standard, using just two straight lines to indicate its shape—perhaps having in mind a post and a beam from which the stars hang down, somewhat like a basketball net. But it is sometimes described as a glittering cluster, and we must remember that, with little or no light pollution, it was surely a rather captivating group of stars hanging off, as it were, the 4th-magnitude star we now call γ Com. If it had not been thought visually striking, its identification would not have been much of a compliment to the queen, or to her predecessor, Arsinoë, from whose temple the West Wind had taken the hair to give to Aphrodite.

Whether or not the hair really disappeared from the temple, or whether the whole episode was fabricated for public consumption, does not matter. ⁷⁸ The point of interest for us is that the poets make the lock speak, rejoicing that it is to be placed in the heavens so that that Ariadne’s Crown (Corona Borealis) will no longer shine “alone.” Significantly, as noted above, Coma Berenices is in the same part of the sky as Corona Borealis, though slightly to the north and further west, with Boötes and the bright star Arcturus between them.

Unfortunately, not only is Callimachus’s text damaged at this point, as elsewhere, but one line in the equivalent passage of Catullus is also “corrupt” (as textual critics say). As with Pliny, different editors reconstruct the texts differently; below I print and translate the version of the relevant lines of Greek that scholarly consensus currently favours, followed by the equivalent lines by Catullus.

In each case I add a translation which seeks to give due relative weight to the different key words and not overlay specific cultural preferences. On the one hand, it is a fallacy to suppose that any translation can ever be “exact,” since all translation is a dialogue between translator and text. On the other hand, one may think of it as analogous to an individual conductor’s interpretation of a symphony: in principle no more and no less “accurate” than any other. A translation can be seen as “the performance of a reading.” ⁷⁹ There is also a distinction between translations that aim to be new works of art with literary merit, free-standing from the original, and those that aim to represent as nearly as possible the cultural presumptions of the original writer in their social and chronological contexts. This interpretative process is one reason why inaccuracy is often demonstrable in the second kind of translation, such as when the translator has been too “free” with the text.

from Callimachus, Aitia, fragment 110 ⁸⁰

  ὄφρα δὲ] ⁸¹ μὴ νύμφης Μινωίδος ο[

60   . . .. . .]ος ἀνθρώποις μοῦνον ἐπι. [ ⁸²

 φάεσ]̣ιν ἐν πολέεσσιν ἀρίθμιος ἀλ̣λ[ὰ φαείνω ⁸³

   καὶ Βερ]ενίκειος καλὸς ἐγὼ πλόκαμ[ος,

 [But in order that] . . . (crown?) of the bride, Minos’s daughter, ⁸⁴ should not . . .

60   [—] for humans alone . . . [—]

 but numbered among many [light]s I shall [show] ⁸⁵

   I, [too,] a lovely Berenicean lock, . . .

from Catullus, Carmina, 66 ⁸⁶

 †hi dii ven ibi† vario ne solum in lumine caeli ⁸⁷

60   ex Ariadnaeis aurea temporibus

 fixa corona foret, sed nos quoque fulgeremus

   devotae flavi verticis exuviae, . . .

 . . . that not . . . alone in the varied light of heaven ⁸⁸

60   a golden crown from the Ariadnean temples ⁸⁹

 should be set, but that we, too, should shine,

   the dedicated spoils of a blond head, . . .

The idea is evidently that Ariadne’s Crown (CrB) will no longer be the only emblem of a royal woman in this region of the sky, for it will be kept company by Berenike’s tresses. The implied comparison of the living Berenice with the legendary Ariadne is doubtless meant to boost the new queen’s (actually somewhat controversial) reputation. ⁹⁰ The new association could even be seen being acted out in the sky: the next two lines in each poem show the lock proudly claiming that it will rise from the sea. Indeed, from Alexandria in late winter and spring, Coma rises from the north-eastern horizon—a sea horizon—immediately before Boötes, with Corona following. ⁹¹ The new queen’s dedication thus visibly takes precedence over the Cretan princess from one of the most famous myths.

The physicist Giorgio Dragoni argues, first, that the particular way in which Callimachus describes the asterism—Ariadne’s Crown will no longer shine “alone”—is strange given that many other easily visible stars lie around, and above all since the very bright Arcturus is directly between Corona and Coma. Accordingly, Conon’s creation of Coma may have been prompted by a celestial event in or near Corona: la “nuova luminosità” vada a tener compagnia a “un’altra luminosità speciale” (“the ‘new brightness’ is going to keep company with ‘another special brightness’”). ⁹² This hypothetical event (so Dragoni reasons) cannot have been a supernova or nova in the modern sense: no filamentary remains have been found telescopically. While it is theoretically possible (he argues) that the event was a comet, the etymological connexion of κομήτης–komētēs with κόμη–komē “hair” suggesting the stratagem to Conon, a comet is typically considered an ill omen and is unlikely to have occasioned the identification of the new catasterism. One might add that associating a comet with just one constellation, when it typically passes through several, would be odd.

Having eliminated comets, Dragoni suggests that a variable star such as T, R, or α Corona Borealis may have been seen at maximum brightness; but he does not develop the argument further. ⁹³ If we do so, we find an interesting outcome. Of these three stars, R is a hopeless candidate: it is normally just above 6th magnitude (the limit of naked-eye visibility), and its defining behaviour is not bursting into brightness but disappearing unpredictably every few years. ⁹⁴ α is no more plausible: its slight variability was discovered only in the early 20th century, ⁹⁵ the change from minimum to maximum on a 17-day cycle being in practice undetectable with the naked eye. ⁹⁶ Only T CrB, among variables in Corona Borealis, is a credible candidate.

On the alternative arithmetical schemes sketched above, T CrB should have been visible, in principle, for a few days about every 80–85 years in the Hellenistic period. Supposing that Conon was motivated by a recent appearance of T which had occurred in 245 bc (year –244) or not long before, this would imply a subsequent eruption in the 160s bc—just at the start of Hipparchus’s career. Thus, instead of attempting to retroject on the basis of (still incompletely understood) astronomical and astrophysical data, we have come at the problem from the other end and employed a “close reading” approach, in order to build a plausible case that teases out hints of meaning from the damaged texts of ancient witnesses who, in any case, are not telling all they know but leaving us to fill in the gaps by sympathetic ratiocination.

Conclusions

In my beginning is my end. We have come full circle from one ancient writer to others, from Pliny to Callimachus and Catullus. As we have seen, Pliny’s testimony favours the sudden appearance of a nova rather than a comet as Hipparchus’s “new star”; his reference to “movement” can be taken to include changes in brightness. Hipparchus’s career as an observer can be extended from 162 to 127/6 bc. The Chinese comet of 134 bc is to be dissociated from Pliny the Elder’s report (and incidentally re-dated to 135 bc, year –134), though a separate attestation of a possible nova in midsummer 134 bc (year –133) raises the possibility that Hipparchus observed a new star in the Libra–Scorpius area; but this is problematically late for him to have only then undertaken his star catalogue. The observed eruptions of T CrB in ad 1946 and 1866 are complemented by reliable reconstructions of eruptions in 1787 and 1217; the date of the latter can now be plausibly pinned down to the last few days of October. Detailed observations of Corona Borealis in Hipparchus’s surviving Commentary, supported by the newly found palimpsest of the relevant section of his lost star catalogue, demonstrate that (as one would have expected) the astronomer paid detailed attention to this constellation, though there is (again as we should expect) no mention of any star we can identify as T CrB. Arithmetical retrojection of the intervals between known and reconstructed medieval–modern eruptions of T CrB, however, is insufficiently secure to predict, even approximately, when the star erupted in the 2nd century bc. Finally, however, there is merit in the suggestion that Conon’s creation of Coma Berenices, in or shortly after 245 bc, may have been motivated by knowing that the nearby constellation of Corona Borealis had witnessed the brief appearance of a new star. If this eruption had occurred recently, it would point to the next eruption falling at the start of Hipparchus’s career.

If T CrB did erupt during Hipparchos’s working lifetime, and was visible to the naked eye for a few days, we can be reasonably confident that he got to see it or knew someone who did. Not only was there far less light pollution in antiquity, but, working at Rhodes and/or Alexandria, his chances of a clear sky were high. For much of humanity today, the phenomenon of T CrB—when, as we expect, it erupts in the near future—is unlikely to be visible owing to light pollution, even given a clear sky. ⁹⁷ Fortunately, the “dark skies” movement is gaining momentum; Greece itself has gained its first International Dark Sky Park, on the island of Kephallenia, joining over 200 others worldwide. ⁹⁸

Unless there are new textual discoveries, we will never be able to prove that Hipparchus saw T Coronae Borealis, or indeed what the “new star” was that he presumably recorded somewhere in his own writings, now lost. Fancifully, one may imagine that he cited it in the introduction to his star catalogue, and that Pliny found the information there. Given the uncertainty, the reader may wonder what value a verdict of “Not Proven” may have. Much of the positive value lies in reflecting upon methodology. In investigating past astronomical events, not only must we be rigorous, rather than unreasonably optimistic, in extrapolating from hard astronomical or astrophysical data, but we must also adopt proper textual-critical and historical-humanistic approaches to written sources. Astronomers must be open to guidance from expert linguists to show the possible interpretations of sources and rule out other interpretations; neither treating texts positivistically nor making unwarranted connexions, but giving due weight to cultural meanings, literary relationships, and historical contexts. We must consider what comets, stars, constellations, and catasterisms meant to our ancestors, and acknowledge that what they thought they saw is what they saw. A catasterism is not a metaphor.