A Probable Flamsteed Observation of the Cassiopeia a Supernova

Baily Francis , An account of the Revd. John Flamsteed … to which is added, his British Catalogue of Stars, Corrected and Enlarged (London, 1835).

Flamsteed John , Historia coelestis Britannica (3 vols, London, 1725). The star catalogue, “Stellarum Inerrantium Catalogus Britannicus”, is in vol. iii, with its own pagination.

Baily discusses the sources of Flamsteed's errors at length in his Account, both in the “Introduction”, especially pp. 370–95, and passim in the notes to the revised catalogue, pp. 506–644.

3 Cassiopeiae is no. 3213 in Baily's corrected edition of Flamsteed's catalogue in the Account, and he discusses the possible identity in a note on p. 641. We will discuss this more fully below.

The catalogue extract is reproduced from Flamsteed's “Catalogus Britannicus”, op. cit., 36. The star map is reproduced from a small section of Plate 15 in Flamsteed John , Atlas coelestis (London, 1753). The plates of this edition are identical to those of the first (1729) edition.

Herschel Caroline , Catalogue of stars, taken from Mr. Flamsteed's observations … and not inserted in the British Catalogue. With an index [and] a collection of errata (London, 1798). The “Index”, in which every star in the 1725 catalogue is cross-referenced to the observations in vol. ii of the Historia coelestis, is still invaluable, provided one remembers that, as in this case, Miss Herschel was not always right.

Ibid., 77. Here, for 3 Cassiopeiae, she references p. 144, line 33, and p. 145, line 2 of the Historia coelestis, vol. ii. Consulting Flamsteed, we find that these refer to two observations of a star “quae praecedit τ”, made on 3 Dec. 1691 at 5^h58^m “circiter”, and on 4 Dec. 1691, at 5^h53^m03^s. Miss Herschel does not reduce the observations; the position I have given is taken from Baily , Account, p. 503, no. 3224. Miss Herschel does say, p. 77, that Flamsteed's catalogue position for 3 Cassiopeiae errs by 5 ¼′ in right ascension and 7′ in polar distance, and although this is consistent with eighteenth-century nomenclature, it can be confusing, because the difference in right ascension is actually 5 ¼ minutes of time or 80 arc minutes.

Baily's discussion is in the Account, note to his no. 3224, p. 642. He points out here that the star observed on 3 and 4 Dec. 1691 was Piazzi XXIII.101, equivalent to (1 Hev.) Cassiopeiae, by which he means that it is the first Cassiopeia star in Flamsteed's version of the Hevelius Catalogue in Historia coelestis, vol. iii. In Hevelius's own catalogue it is actually the 30th star in Cassiopeia. Because of this confusion, I have somewhat anachronistically used the modern designation, AR Cassiopeiae. The reduction by Flamsteed is extant in MS xxiii, 360–3, according to Baily.

Ibid., 641, note to no. 3213; the comment by Flamsteed is in MS xxviB, 36, according to Baily.

Baily's final comment was that 3 Cassiopeiae “must have been introduced through some mistake in the trigonometrical computation, as there is no star to be found corresponding with the position here given”, idem. The “singular” remark is in the note to no. 3224, p. 642.

Ryle M. and Smith F. G. , “A new intense source of radio-frequency radiation in the constellation of Cassiopeia”, Nature, clxii (1948), 462–3. The authors' initial position for the source was later refined, and the position I have given is that determined by Rosenberg in 1970 and related by van den Bergh and Dodd (see ref. 13, below).

Baade W. and Minkowski R. , “Identification of the radio sources in Cassiopeia, Cygnus A, and Puppis A”, Astrophysical journal, cxix (1954), 206–14, including some fine 200 plates of the nebulosity. The realization that Cassiopeia A was a supernova remnant came in later papers.

van den Bergh Sidney and Dodd W. W. , “Optical studies of Cassiopeia A. I: Proper motions in the optical remnant”, Astrophysical journal, clxii (1970), 485–93. The authors derived a position for the centre of expansion of α₁₉₅₀ = 23^h21^m11^s, δ₁₉₅₀ = + 58°321919”, which differs slightly from the centre of the radio source; however, differences of a few seconds may be disregarded here, as we shall see.

Van den Bergh and Dodd , ibid., discuss the lack of seventeenth-century observations and conclude that due to absorption and a low maximum brightness (Cassiopeia A is some 2·8 kpc distant), the absence of observations is not surprising. Chevalier Roger A. , in “Cassiopeia A, faint supernovae, and heavy-element ejection by supernovae”, Astrophysical journal, ccviii (1976), 826–8, maintains that the lack of observations is significant and means that Cassiopeia A was not a normal supernova. Shklovsky I. S. , “Is Cassiopeia A a black hole?”, Nature, cclxxix (1979), 703, agrees, although he postulates a different order of events, as indicated by his title.

All calculated positions and angular distances in this article are correct only to the nearest tenth of a minute. Further precision seems unnecessary in light of the discrepancies of 6 arc minutes which we must confront. For this same reason, while the effects of precession and refraction have been incorporated to the degree of accuracy necessary, the effects of nutation and aberration have been ignored. However, whenever a position or measurement is being transcribed, whether from Flamsteed, Baily or modern authorities, I have reproduced the figures exactly as given.

Baily Francis (ed.), The catalogue of the British Association for the Advancement of Science (London, 1845).

See ref. 20 below.

Flamsteed John , Historiae coelestis libri duo, ed. by Halley Edmond (London, 1712); Historia coelestis (1725), vol. i. The word “reappeared” is quite appropriate here. In view of the often-repeated story that Flamsteed bought up the remaining copies of the 1712 edition and burned them, it is perhaps worth reiterating that (a) he did not buy them at all—they were given to him, and (b) he destroyed only the introductory and concluding material and the catalogue. The rest, 387 pages in all, were retained and inserted intact into vol. i of the 1725 edition, and these pages include the sextant observations of the fixed stars.

The extract is reproduced from p. 63 of the “Observationes Fixarum” section of the 1712 edition of the Historia coelestis; it may be found on the same page of the same section in the 1725 edition, vol. i.

The approximate positions of “supra τ” and the determining stars at the time of observation are given in Figure 5. Flamsteed's catalogue position for β Persei is α₁₆₉₀ = 42°21915”, δ₁₆₉₀ = + 39°43′10”; for β Pegasi, α₁₆₉₀ = 342°11′25”, δ₁₆₉₀ = + 26°24′20”. Disregarding refraction, the “trigonometric process” (as Baily liked to call it) yields a position for “supra τ” of 347°38′·7, + 56°57′·9. Allowing for refraction (and note that β Persei had an altitude of only 23° when the sextant measurement was taken), we obtain a position of 347°35′·6, + 56°58′·7. Since the catalogue position of 3 Cassiopeiae (347°37′0”, + 56°58′30”) lies between these two positions, it seems safe to conclude that (a) 3 Cassiopeiae is the reduced position of “supra τ”, and (b) Flamsteed underestimated the amount of refraction. Baily; Account, in his note to 3 Cassiopeiae, wondered if the position might have derived from the sextant observation of “supra τ”. However, he pursued the point no further, and his reference to vol. ii, p. 63, instead of vol. i, p. 63, ensured that few were going to follow in his footsteps.

The revised positions for the determining stars were taken from Baily , Account, Table K, 656–66, where he provides corrected positions for 345 determining stars to epoch 1690.

The sextant readings should have been 30°55′·4 and 38°51′·0, rather than the readings Flamsteed obtained of 30°48′·4 (30°49′·4 by the screw reading) and 38°45′·1.

It is important to note that we are concerned here only with observational error. Flamsteed regularly made sizeable errors by miscopying numbers or performing incorrect calculations; Eric Forbes, in the Introduction to his edition of The Gresham Lectures of John Flamsteed (London, 1975), 20–27, gives a compelling demonstration of how Flamsteed managed to produce cometary positions that were off by several degrees, through this kind of error. Baily, in his Account, offers many other examples of calculatory errors. But while these kinds of miscalculations provide ready illustrations of what can go wrong in astronomy, it would be misleading to introduce them here. There is no problem in the calculations that will explain the discrepancy.

See Flamsteed's “History of his own life”, in Baily , Account, 38, and Forbes Eric G. , “John Flamsteed and the origins of the Greenwich astronomical tradition”, Journal of navigation, xxviii (1975), 251–62, p. 256.

The determinations are Baily's; see the corrected Catalogue in Account, nos. 2112, 2625, and 1579 respectively, and the notes thereto.

Baily , in fact, in Account, Table G, 651–3, lists 115 stars where the right ascension is in error by more than 4′, and these do not include calculating errors, which Baily corrects in the catalogue itself.

Baily , Catalogue of the British Association , 76. This same table appeared in Account, Table C, 646, but there Baily included uncatalogued stars and the number was swollen to 23. It is perhaps worth mentioning that much of the introductory material in the Account was reproduced nearly verbatim in the introduction to the BAAS Catalogue.

As a further reminder that Flamsteed was not in the habit of observing non-existent stars, it is instructive to remember that on 13 December 1690 Flamsteed observed a star in Taurus which also later turned out to be missing, and he subsequently found five similar stars in Leo in 1712 and 1715 which no one else could find. All of them later turned out to be observations of a very real object—the planet Uranus. See Baily , Account, 393.

[ Bevis John , Uranographia (London, c. 1750).] The work was never published, because of financial difficulties, and only sets of proof impressions survive. The so-called Atlas celeste of 1786 is a ghost, consisting of original proof impressions with a bookseller's advertisement bound in as a title page.

The ecliptic coordinates of Cassiopeia A, epoch 1690, are λ₁₆₉₀ = γ23°6′·4, β₁₆₉₀ = + 54°49′·5. Bevis obtained the 1750 positions of his stars by simply adding 50′·3 of longitude to Flamsteed's values. If we do this for Cassiopeia A, we obtain λ₁₇₅₀ = γ23°56′·7, β₁₇₆₀ = + 54°49′·5. This puts Cassiopeia A about 11′ due north (in an ecliptic sense) of 3 Cassiopeiae, whose 1750 position was, according to Bevis, λ₁₇₅₀ γ23°55′·4, β₁₇₅₀ = + 54°38′·5. The identification of the stars on the Bevis plates is made possible by the recent discovery of Bevis's star catalogue, long thought never to have been printed. The author has an article forthcoming on the Bevis atlas and the unique catalogue, which may be found in the Library of the American Philosophical Society, Philadelphia. Although the catalogue is for epoch 1750, the star maps are actually for epoch 1746. The difference is, however, not easily detectable on the printed plates, and certainly does not affect our comparison of the positions of Cassiopeia A and 3 Cassiopeiae.

Clark David H. and Stephenson Richard F. , in The historical supernovae (Oxford, 1977), 207, arrive at a tally of four certain, two probable, and two possible historical observations of galactic supernovae. Flamsteed's observation becomes the ninth and the latest.