8=15 UC-NRLF SC 34 R PHYSICS COMMITTEE. I OF T ETY UF THE PACIFIC ON SOME OF THE PHENOMENA OF NEW STARS. LONDON : PRINTED UNDER THE AUTHORITY OF HIS MAJESTY'S STATIONERY OFFICE BY EYRE AND SPOTTISWOODE, LTD., EAST HAKIMXU STIJEJ PRIXTEHS TO THE KIX<:'? MUST EXCELLENT MAJF- To be purchased, either directly or through any Boo' oin WYMAN AND SONS, LTD., 29, HKEAMS Bi ILDINGS." FETTER LAS and 54, ST. MAI;Y STHKET. CAKDIFF : or H.M. STATIONERY OFFICE (SCOTTISH BKAXCH), 23. F.HITH -s EI.IXBI H..II : or E. PONSONBY, LTI... 116, QUAFTOX STREET. Driu or from the Agencies in the British Colonies and Depen< the United States of America, the Continent of Europe and Atv-oad of T. FISHER UNWIN. LONDON. W.C. 1914. Fire Shillings. GOVEENMENT PUBLICATIONS (with the under- mentioned exceptions) can be purchased in the manner indicated on the first page of this wrapper. Booksellers, and the accredited agents of Free Public Libraries, are entitled to a discount of 25 per cent, from published prices. Hydrographical Publications of the Admiralty are sold by J. D. POTTER, 145, Minories, London, E.G. Patent Office Publications are sold at The Patent Office, 25, Southampton Buildings, Chancery Lane, London, W.C. (N.B. Classified Abridgments of Patent Specifications are sold also by Wyman and Sons, Ltd.) Ordnance Survey and Geological Survey Publications can be purchased from The Director General of the Ordnance Survey, Southampton ; or The Superintendent, Ordnance Survey, Dublin ; or Agents in most of the chief towns in the United Kingdom. (N.B. Small Scale Slaps are, as a rule, procurable at Railway Bookstalls in England and Wales.) The Journal of the Board of Agriculture is published monthly by the Board, at 4, Whitehall Place, London, S.W. Price 4d., post free. The London Gazette is published on Tuesday and Friday evenings by Wyman and Sons, Ltd. Price Is. The following is a list of some Official Publications : SOLAR PHYSICS COMMITTEE. AUSTRALIAN METEOROLOGY. A Discussion of Australian Meteorology. Being a Study of the Pressure, Rainfall, and River Changes, both Seasonal and from Year to Year ; with a Comparison of the Air Move- ments over Australia with those over South Africa and South America. (1909.) Price 5s., post free 5s. 5d. BAROMETRIC PRESSURE. Monthly mean Values of Barometric Pressure. For 73 Selected Stations over the Earth's surface. Being a Comparison of Barometric Variations of Short Duration at places widely distributed. (1908.) Price 6s., post free 6s. 5d. BAROMETRIC PRESSURE AND RAINFALL. Mean Annual Variation of Barometric Pressure and Rainfall in Certain Regions. (1905.) Price Is. Qd., post free Is. 8d. BRIGHTER STARS. Catalogue of Four Hundred and Seventy of the Brighter Stars. Classified, according to their Chemistry, at the Solar Physics Observatory. (1902.) Price 5s. Qd., post free 5s. Qd. GENERAL SPECTRA OF CERTAIN TYPE-STARS. General Spectra of Certain Type-Stars and the Spectra of several of the Brighter Stars in. the Green Region. (1908.) Price 3s., post free 3s. 3d. SOLAR ECLIPSE EXPEDITIONS. Report of the Solar Eclipse Expedition to Palma, Majorca, Aug. 30, 1905. (1907.) Price 3s. 6d., post free 3s. 10d. Report of the Solar Eclipse Expedition to Vavau, Tonga Islands, April 29, 1911 (Eastern Date). (1912.) Price 6s., post free 6s. 4d. SOLAR AND STELLAR SPECTRA. Researches on the Chemical Origin of Various Lines in Solar and Stellar Spectra. (1910.) Price 4s. 6d., post free 4s. lOd. SPECTRA OF SUN SPOTS, 1879-1897. Spectra of Sun Spots, 1 879-1897, deduced from Observations made at the Solar Physics Observatory, South Kensington. (Printed in 1900.) (1913.) Price 7s. Gd., post free 7s. Qd. T J GIFT OF ASTRONOMICAL SOCIETY OP ' PACIFIC ^ SOLAR PHYSICS COMMITTEE. ON SOME OF THE PHENOMENA OF NEW STARS. UNDER THE DIRECTION OF Sir NORMAN LOCKYER, K.C.B., LL.D., D.Sc., F.R.S. LONDON: PRINTED UNDER THE AUTHORITY OF HIS MAJESTY'S STATIONERY OFFICE BY EYRE AND SPOTTISWOODE, LTD., EAST HAKDIXG STREET, B.C., PRINTERS TO THE KING'S MOST EXCELLENT MAJESTY. To be purchased, either directly or through any Bookseller, from VVYMAN AND SONS, LTD., 29, BREAMS BUILDINGS, FETTER LANE, B.C., and 54, ST. MARY STREET, CARDIFF ; or H.M. STATIONERY OFFICE (SCOTTISH BRANCH), 23, FORTH STREET, EDINBURGH ; or E. PONSONBY, LTD., 116, GRAFTON STREET, DUBLIN; or from the Agencies in the British Colonies and Dependencies, the United States of America, the Continent of Europe and Abroad of T. FISHER UNWIN, LONDON, W.C. 1914. Price Fire ShilUnys. ASTRONOMY DEFT. uu u AcUon. Dipt LIST OP PUBLICATIONS OF THE SOLAE PHYSICS COMMITTEE FROM 1879 TO 1914. Year of Publication. Preliminary Report of the Committee on Solar Physics (1880) Report by the Committee on Solar Physics - - (1882) Second Report of the Committee on Solar Physics - - (1889) Positions and Areas of Sim Spo'ts and Faculie, 1887-88 - (1889) 1889 (1891) 1878-1881 (inclusive) and 1890 (1892) ,. 1891 - (1894) . ,, 1892 - (1895) 1893- (1896) 1894, 1895, 1896 and 1897 (1899) Spectra of Suu Spots 1879-1897 (inclusive) (1900) Positions and Areas of Sun Spots and Faculfe, 1898 and 1899 - (1.901) Catalogue of 470 of the Brighter Stars Classified according to their Chemistry - (1902) The Sun's Spotted Area, 1832-1900 (1902) Positions and Areas of Sun Spots and Faculfe, 1900, 1901 and 1902 (1904) Mean Annual Variations of Barometric Pressure and Rainfall - (1905) Positions and Areas of Sun Spots and Facuke, 1903 and 1904 - (1906) Tables of Wave-lengths of Enhanced Lines - - (1906) Spectnoscopic Comparison of Metals present in certain Terrestrial and Celestial Light- sources (with special reference to Vanadium and Titanium) - (1907) Report of the Solar Eclipse Expedition to Palma, Majorca, August 30, 1905 - (1907) Positions and Areas of Sun Spots and Faculie, 1905 - (1907) Monthly Mean Values of Barometric Pressure for 73 Selected Stations over the Earth's Surface (1908) On the General Spectra of certain Type-Stars and the Spectra of several of the Brighter Stars in the Green Region - (1908) Positions and Areas of Sun Spots and Facula;, 1906 - (1909) A Discussion of Australian Meteorology - - (1909) Southern Hemisphere Surface Air Circulation - (1910) Researches on the Chemical Origin of Various Lines in Solar and Stellar Spectra - - (1910) Positions and Areas of Sun Spots and Facula?, 1907 and 1908 - (191 1) v 1909 - (1911) >, ., ,, v 1910 (1911) Report of the Solar Eclipse Expedition to Vavau. Tonga Islands. April 29. 1911 (Eastern Date) (1912) (I) On the Spectra of the Rigelian, Crucian and Alnitamian Stars ; (II) The Line Spectrum of a Orionis ; (III) The Spectrum of y Cassiopeia (one volume) - - (1914) t Areas of Calcium Flocculi on Spectroheliograms, 1906-08 - - (1914) On Some of the Phenomena of New Stars - - (1914) ** *, . **** CONTENTS. PAGE 1. On the Spectra of New Stars : Introduction - - 1 Novae discovered since 1891 (Table I.) - 1 2. The Spectra of the Fainter Nova? - 2 The Spectra of the Fainter Novas (Table II.) - - - - 3 3. The Spectra of the Brighter Novas : (a) The sequence of the Appearance of different Types of Lines - 4 (6) The Continuous Spectrum (with or without Dark Lines) first Observed - - 6 Comparison of the " Dark-line " Spectrum of Nova Persei (Pickering) with the Spectra of Type-Stars (South Kensington) (Table IV.) - - - - -9 The Dark Line Spectrum of Nova Geminorum (2) - - - - 13 (e) A Comparison of the Dark-line Spectrum of.Novas and the Spectra of Globular Clusters 14 (d) The "Bright-line" Spectrum - - - - - 15 (e) Principles employed in the Reduction of Nova; Wave-lengths at South Kensington - 17 / (./") Comparison of the Bright Lines in the Spectra of the Second Stage of various Novas with the Strong Absorption Lines of the a. Cygni Spectrum (see Table V.) - - 18 (g) Special Study of the Stage producing the Bright Band near X 4640, in the Spectra of Novae - . - - - 21 (A) The " Nebula Spectrum " of Novse - ... 27 (/) On the Spectrum of Nova Geminorum (2) : Preliminary General Description - - - - - . -28 Nova Geminorum (2) wave-lengths, March 19, 1912 (Table VIII.) " Bright-line stage" . - .30 Nova Geminorum (2), wave-lengths April 7, 1912, " 4640 stage" (Table IX.) - 31 Band near X 4640, exact wave-length (Table X.) - - - - - 32 Detailed discussion of the Spectrum of Nova Geminorum (2) March 16-April 16, 1912 . - - 33 4. Relation of the Novas Spectra to the Bright-line Spectra of other Stars : (a) Wolf-Rayet Stars - > - - - - 41 Wolf-Rayet Stars, and Novas in Metallic Bright-line Stage (Table XI.) - - 42 " Nebula " Stage (Table XII.) - 47 Spectra of N. Geminorum (2) in X 4640 Stage and the Type Star of the Oa spectral class (Table XIII.) - ........ 52 (b) Other Bright-line Stars and Novas (Table XIV.) - - - 54 5. The Structure of the Bands in Novas Spectra - - - 56 6. Magnitude Observations of Novse - - 58 7- Location of Novas in Space - ... ...59 ADDENDUM. Adams and Kohlschutter's published reduction of the Spectrum of Nova Geminorum No. 2 - 60 a (11)19881 Wi 2807272 375 4/14 E & S A. 2 701223 LIST OF ILLUSTRATIONS. PAGE Fig. 1. Typical Spectra at various stages - 5 2. The Stellar classification - 8 Plate I. Comparison of Nova Geminorum (2) ; .Spectrum with the typical spectra of the ) Cvgnian and Rigelian groups (South Kensington) (sec page 8) - At end Plate II. Nova Geininornm (2) ; Madrid spectra (see page 33) - J> ( >f Plate III. Comparison of Light-Curves (see page 59) - Plate IV. Relation of Novas to Milky Way (see page 60) PREFACE. The materials for this memoir have been accumulating for many years, indeed since my papers on Nova Persei communicated to the Royal Society in 1901-02. Mr. Rolston has recently been employed on them, under my supervision, because I was anxious to bring them together before my work here ceased. Very fortunately, while this was going on, Professor Iniguez, of the Astro- nomical and Meteorological Observatory, Madrid, was good enough to send me a most valuable series of photographs of the spectrum of the recent nova, taken by him, with full permission to reduce them. This reduction has been made by Mr. Baxandall with great care, and I think it is clear that the Madrid results afford a most valuable comparison with, and confirmation of, those obtained from other novae. They also enable the sequence of the phenomena to be studied with greater precision than was possible before. It was my original intention that this memoir should be more comprehensive than it is, e.g., there is the large question of the nebulosity connected with novae, as in the case of Nova Persei (2), but the pressure of work incident upon the removal of the Observatory has made a complete discussion impossible. The general conclusions to be derived from the collection of data now brought together must therefore be deferred to a further memoir. NORMAN LOCKYER. Solar Physics Observatory, South Kensington. March 17, 1913. 1. ON THE SPECTEA OF NEW STARS. Introduction. This memoir is a continuation of that which I communicated to the Royal Society in 1890 :!: " on the causes which produce the phenomena of new stars. Up to that time the observations of novae had been visual. In the first memoir an account was given of the various theories from those of Tycho Brahe and Kepler, which had been put forward to explain the sudden increase of light. I was the first in 1877 to ascribe it to meteoric collisions,! and the fading away to a backwardation due to a reduction of temperature of the phenomena seen in stars increasing their temperature. This memoir also contained and discussed the eye observations made up to that time. In the present memoir I propose to discuss the spectra obtained by the photographs taken in later years, the first dating from 1892, when a nova made its appearance in the constellation Auriga. The following table gives the names and the dates of discovery of the various objects. In the first column are given the dates of discovery of the nova named in the second column. In the third, fourth, and fifth columns the dates of the first observation are recorded and show in each case whether the observation was made visually, photographically, or spectroscopically. The sixth column gives the date that the apparition of the nova was discovered, and except in the case of the visual observations this need not necessarily coincide with the date of the first observation, for in the case of the photo- graphic observations, direct or spectroscopic, the plate was sometimes not examined until some months after it was taken ; the name of the discoverer is given in the seventh column. The eighth column gives the magnitude at the time of discovery. TABLE I. PARTICULARS CONCERNING THE XOVJE DISCOVERED SINCE 1891. 1 t Year R.A. Dec. Date first observed Magni- K Xame. announced. (1900.) (1900.) Photo- Visually. graphi- cally. Spectro- scopically. Discoverer. Time of Discovery. H. M. <, l 1892 K. Aurigae a 25-6 30 22 N. 1.2.92 10.12.91 Anderson - 4-4 2 1893 X. Normae - 15 22-2 50 14 S. 10.7.93 Fleming - 7'0 3 1895 N. Centauri 13 34-3 31 8 S. 8.7.95 18.7.95 Fleming - 7 "2 4 189fi N. Carinae - 11 39 61 24 S. 8.4.95 14.4.95 Fleming - 8'0 i ' Phil. Trans., 182 A , p. 397. t Nature, 16, p. 413 : reprinted Phil. Trans., 182, p. 405. *"* ''* SOLAR PHYSICS COMMITTEE. PARTICULARS CONCERNING THE NOV^E DISCOVERED SINCE 1891 continued. Number. Y ear announced. Name. R.A. f!900.) Dec. (1900) Date first observed. Discoverer. Magni- tude at Time of Discovery. Visually. Photo graphi- cally. Spectro- scopieally. H, 11. , 1 5 1899 N. Sagittarii (1) - 18 56-2 13 18 S. 8.3.98 19.4.98 Fleming 4-7 6 1900 N. Aquilas (1) - 19 15-3 19 S. 21.4.99 3.7.99 Fleming 7-0 7 1901 N. Persei (2) 3 24-4 43 34 N. 22.2.1901 22.2.1901 22.2.1901 Anderson - 2-7 8 1903 N. Geminorum (1) 6 37-8 30 3 N. 16.3.03 26.3.03 Turner 8-0 9 1905 N. Ophiuchi (3) - 17 44-8 16 40 S. 1888 15.7.98 Fleming 7-7 l(t 1905 N. Aquilae (2) - 18 56-8 4 35 S. 18.8.05 Fleming 9-0 11 1906 N. Scorpii (2) - 17 47-5 34 20 S. 14.6.06 Cannon . 8-8 12 1906 N. Velorum 10 58-3 53 51 S. 5.12.05 Leavitt 9-7 18 19u7 N. Circini - 14 40-5 59 35 S. .2.06 Leavitt 9-5 14 1910 N. Arse Ifi 33-0 52 14 S. 4.4.10 5.7.10 Fleming 6-0 15 1910 N. Sagittarii (2) - 17 53-8 27 33 8. 31.5.10 21.3.10 Fleming 7-8 16 1910 N. Sagittarii (3) - 18 13'8 25 14 S. 10.8.99 Cannon 8-5 IT 1910 N. Lacertae - 22 31-8 52 12 N. 30.12.10 19.11.10 1.1.11 Espin 8-0 18 1'Jll N. Sagittarii (4) - 18 0-5 27 26 S. 22.5.01 Cannon 10-3 19 1912 N. Geminorum (2) 6 48-4 32 16 N. 12.2.12 13.2.12 Enebo . 4-5 In the case of a number of these novae, the star was too faint when discovered for its spectrum to be secured, in other cases the spectrum which revealed its nova character was insufficient to give much more information ; it will therefore be impossible to include the objects in the discussion. On these grounds Nos. 11, 13, 16, 17 may be dismissed at once. 2 THE SPECTRA OF THE FAINTER We have some spectroscopic data for N. Aurigae (1892), N. Normae (1893), N. Carinaj (1896), N. Centauri (1895), N. Sagittarii (1) (1899), N. Ophiuchi (1898), N. Aquite (1) (1899), N. Persei' (2) (1901), N. Geminorum (1) (1903), N. Aquilse (2) (1905), N. Velorum (1906), N. Sagittarii (2) (1910), N. Arae (1910), N. Lacertaa (1910), and IS. Geminorum (2) (1912). In four cases only, viz., N. Auriga?, N. Persei, N. Lacertae and N. Geminorum (2) are the data full enough to warrant detailed discussion, and these four will be discussed later. N. Centauri, although providing meagre data, appears to have presented special features calling for further remark, but for the present we will dispose of the remaining ten by tabulating the lines recorded in their spectra. PHENOMENA OF NEW STARS. TABLE II. SPECTRA OF THE FAINTER XOV^E. Nova Normse. Nova Cannae. Nova Nova Ophiuchi Sagittarii ( 2 ). (R.S. O)- Ophiuchi). Nova /ova Aquil*. 1 Gemi :, . norum (1). i Nova Nova Aquilae (2). Velorum. Nova Saaittarii (2)- Nova Arse. 1894, Feb. 13. 1895, 1898, 1898, April 14. April 19. July 15. 1899, 1903. 1903, - 1906. 1910, July 3.|| April 2. Aug. 18.^ June 6. 1 Mar.-Juue. 1910, July 5. AA 335 350 374 H, 384 383 383 Hf H? Hf (2) Hf 389 3889 3889 H, H H. (1) H 397 3970 3970 4029 Hi Hi Hi (5) Hi 410 4101 4101 4101 410 4179 4238 4276 H, Hy Hy (10) Hy 434 4341 4340 4341 4341 4459 446 4472 4530 4611 _ (4700)t 4643 4656(7) 462 4646 4643 4670 _ 4691(14) 4693 486 H/i Hp H0 (20) H0 486 4861 4862 4861 4861 4926 4926 4953 * J(5005) : 5007 5003 5007 5007 L*5013 5013 5175 Antrim uitd Antro- PttVs'K-*. Vol. XIII.. p. 311 (Camp- bell). Harv. Coll. Obs.. Circular No. 1. Harv. ColL Obs.. Circular No. 42. Harv. Coll. Obs., Circular No. 76. Harv. Coll. Obs., Circular No. 56 (Pickering). Lick Obs., Bulletin No. 37, p. 298 (Perrine). Harv. Coll. Obs.. Circular No. 106. Harv. Coll. Obs., Circular No. 121. Aitro- Homitche JTacfiricft- ten. No. 4448. Harv. Coll. Obs. Annals, Vol. LVL, Part VI., p. 170. * Brightest line in the spectrum on March 6. 1894 ; measured wave-length 5007-3. t Line about A 4700 barely visible on April 14 was as bright as hydrogen lines on June 15. continuous spectrum very faint. t 5005 appeared on April 21. 1898. Variable. Observed at Harvard since 1888. On August 28, 1894. spectrum was K (Arcturian-Aldebarian) type without trace of bright lines. || September 7, 1899. H y and \ 4959 ? and on October 27, 1899, H y and A. 5007 alone visible. ^ Spectrum like that of Nova Persei on March 30, 1901. (Harv. Coll. Obs. Annals, Vol. LVL, Part VI., p. 170.) SOLAR PHYSICS COMMITTEE. Of the nova which appeared in the constellation Centaurus in 1895, Pickering stated* that the spectrum on July 18, 1895, resemhled that of the nebula surrounding 30 Doradus, and was unlike that of an ordinary nebula or of the new stars in Auriga, Norma and Carina. According to Le Sueurf, the spectrum of 30 Doradus, examined with the great Melbourne reflector, showed the first nebular line 500 easily and the 497 line certainly but with difficulty, while the hydrogen line F was only suspected. On December 16 and 19, continues Pickering, Mr. Wendell found the spectrum of the nova to be monochromatic and very similar to the adjacent nebula N.G.C. 5253. The spectrum of this nova was examined by Prof. Campbell at Mount Hamilton on December 22 and 29, 1895, and seemed to himij: to be continuous, certainly not nebular. Re-examining it on February 8, 1896, under much better conditions, he found that the spectrum of the nova certainly was continuous although very peculiar. The maximum visual intensity was in the yellow-green, the green-blue was very faint, while the blue Avas surprisingly strong, in fact was very much brighter visually than the green-blue, and the spectrum was relatively faint about X 5200 to X 4600. There was no trace of the nebular lines or of the H0 line, although there was some evidence of bright lines or of irregularities in the brightest portion of the spectrum, but the light was too weak to enable this to be established with certainty. Prof. Campbell suggests the possibility that under the difficult conditions of the Harvard observations the spectrum of the neighbouring nebula, in which he finds the nebular lines with their usual intensities, was observed. While there was a difference of opinion as to the inclusion of this object as a nova, it is now generally accepted as such, and we must consider it as a new star which exhibited a continuous spectrum. Subsequent observations by Prof. Hussey, quoted by Campbell, showed the star to be surrounded by a faint irregular nebulosity which was joined up to the adjacent large nebula N.G.C. 5253 ; this was on July 9, 1896, when the nova's magnitude was about 16. 3 THE SPECTRA OP THE BRIGHTER NOV^E. (a). THE SEQUENCE OF THE APPEARANCE OF DIFFERENT TYPES OF LINES. We are now left to deal with the four objects named in the preceding section which have been sufficiently bright to afford spectra from which the information concerning the physical conditions of the radiating and absorbing matter involved could be deduced. * Harvard College Observatory Circular No, 4. t Proc. Roy. Soc., Vol. 18, p. 222. . 190;) Lacertie, 1910 - - Dec. 80. 11)10 Jan. 2. 11)11 Geminorum (2). 11)12 Mar. 12, 11)12 Mar, H. 1912 (Ediiiboro'). (Harvard). (Harvard). (South Kensington). Mar. 28, 1903 1 (Frost). (Lick). Jan. 3, 1911 Mar. 30. 1911 (Wright). Mar.13,14,1912 Mar. 15, 1912 Mar. 27, 1912 April (i. 1912 (Harvard). (Harvard). (Madrid). (Mt. Wilson). NOTK. These dates, and therefore the intervals, are by no means rigid, because, in the first place, the change is a gradual one. difficult to determine, ami, secondly, the time when such a change was first observed was not necessarily the actual time of its inception. 3 (i). THE CONTINUOUS SPECTRUM, WITH OR WITHOUT DARK LINES, FIRST OBSERVED. In 1901, when Nova Persei (2) appeared, Dr. Copeland examined the spectrum ahout 16 hours after the discovery of the nova, and found it to he strong and purely continuous."* When the discovery was generally known and weather permitted observations to he made this phase of the spectrum had disappeared, and, so far as we know, it has not been observed in the case of any other nova with the exception of the peculiar Nova Centauri referred to in the previous section. In the case of Nova Persei (2) dark absorption lines appeared within a few hours in the continuous spectrum. Of the novae discoA'ered since 1891 it has been definitely recorded of two only that a purely dark (absorption) line spectrum stage has preceded that stage in the spectral changes in which the spectrum is essentially one of bright lines or bands, and is generally known as the "typical nova" spectrum. The two stars are Nova Persei (2) and Nova Geminorum (2). Pro. Roy. Soc., Edinburgh, Vol. XXV., Part I., p. 545, 1. PHENOMENA OF NEW STARS. The small proportion of novae in which this stage has l)een observed is not surprising, for even in the two cases named the condition was remarkably transient, lasting but a few hours,* and those few hours would occur when observers were likely to be in a state of hurried preparation. Further, many of the remaining nova; were not recognised as such until months after they had been recorded by photography, and the Harvard plates on which they were recorded are of too small a dispersion to give any indications on such a fine point.t Possibly such a stage was recorded by Huggins in the nova of 18GG, Xova Corona;, but his statement* is not very definite. In the first place, the absorption spectrum was contemporaneous with the bright line spectrum and therefore is on a different basis to the purely absorption spectra recorded by Pickering in the later nova'. The statement made by Huggins goes no further than to say that in general features, i.e., in the breadth of the absorption lines and in the absence of hydrogen absorption, the spectrum resembles a Orionis (Antarian). Such a statement made in the early infancy of spectroscopy affords too weak a basis for us to class the Huggins absorption spectrum of Xova Corona 1 with those purely dark-line spectra observed in the later nova 1 by Prof. Pickering. Keverting to the consideration of these, we find that in the case of Xova Persei (2), Pickering described the spectrum in general terms as similar to that of Rigel,$ and that of Xova Geminorum (2) as similar to Procyon. As these are two type-stars in the Kensington classification it will be convenient here to show their position in that classification where an attempt has l>een made to trace out, more or less consecutively, the life history of the stars, not merely to pigeon- hole them with labels having but little reference to their state of condensation or any scheme of evolution. Thus, we consider the Antarian class (e.g., a Orionis) as the earliest known form of star, which later, by contraction, becomes hotter and hotter until the Argonian stage is reached, and then by excess of radiation, cools down through the various stages, including that to which the sun belongs, to the oldest known stellar type, the Piscian. Our diagram I Fig. 2) shows this graphically, as well as the change of condensation and the different conditions on each arm of the temperature curve. For each stage the Harvard designations of Miss Maury (M) and Miss Cannon (C) are given above and below, respectively, the Kensington name. ' Harvard Coll. Olis. Annals Vol. LVI., No. III., p. 44. t See Prof. E. C. Pickering, Ast. Xa.-li. I Proc. Roy. Soc., Vol. XV.. j>. 147. 18(30. Harvard Coll. Obs. Annals, Vol. LVI.. Xo. III.. j>. 41. i| Harvard Coil. Ohs. Giro. Xo. 17G, p. '2. P, 2 SOLAR PHYSICS COMMITTEE. PIG. 2. THE STELLAR CLASSIFICATION. In the accompanying table (IV.) the dark lines observed photographically in N. Persei, at the Harvard College Observatory, on February 22, 1901, are com- pared with the dark lines of the type-stars Higel, Bellatrix, Markab, and a Cygni ; where two wave-lengths are given without a bracket, the lower one represents a line in a |_S. Kensington type-star. Pickering gave the spectrum at this time as similar to that of Eigel. Sec Plate I. The following analysis of the lines in detail will enable iis to study this question : The first two lines common to Nova Persei and the other spectra are due to hydrogen, and need no further remark. 4026 '4 (3) Helium line. Comparatively weak in the nova (3 on a maximum of 40), and is therefore more like Markab (intens. 1-2 on maximum of 10) or a Cygni (intens. 1 on maximum of 10). 4076 "9 (2) In Eigel and Bellatrix this is probably due to oxygen, the other (407(5-3) oxygen lines near by being present. But in the nova this is the only oxygen line of the group that is possibly represented, and its oxygen origin seems to be unlikely. PHENOMENA OF NEW STAKS. TABLE IV. COMPARISON OF THE "DARK LINE" SPECTRL'M OF NOVA PERSEI GIVEN BY PICKERING WITH LINES IN THE SPECTRA OF S. KENSINGTON TYPE STARS. Nova Persei. Pickering's Absorption Spectrum. Feb. 22, 19ul. Rigel. S. Kensington Bellatrix, Markab. Cygni, S. Kensington. S. Kensington. S. Kensington. Origins in ! Tvpe- A A Xe.'ative(l) A A Negative (2) Int. Max i 1"'- j = 40. CO) T Stars. Int. Int. Int. (10) (Id) (Hi) 3893-6 15 3889-1 10 3889-1 10 3889-1 8 3889-1 10 H, 3970-2 3970-2 25 3970-2 10 3970-2 10 3969-5 10 3970-2 10 H, 4026-4 4026-4 3 4026-3 5-6 4026-3 8 4026-3 1-2 4026-3 1 He 4076-9 2 4076-3 < 4076-3 2 - 4101-7 ( 4093-8. ' 4102-8 ) (4107-3/ 30 4101-8 10 4101-8 4077 9 1 . 10 4101-8 10 40779 3 p. Sr 410T8 10 H, 4126-4 j 4128-1 5 4129-1 1 4128-1 2 4128-1 5-6 p. Si (II.) 4151M 4159-1 1 ( 4153-0 1-2 4154-0 2 O 4266-4 4266-1 2 4267-5 3- 4 4267-5 3 4267-5 < :i p. c. 4340-7 , 4334-1, 4340-7) ( 4346- 3 / 40 4340-7 1 4340-7 10 4340 7 10 4340-7 10 H T 4365-5 4366 4 1 4365-4 < a p. MI. :- 4388-4 4388-6 3 4388-1 3-4 4388-1 6 4388-1 1 Ast. 4415-3 1 4415-7' *) L 4415-3 < :l Fe 4417-0 1 4417-2* ., ( 4417-0 ) i 4417-0 ' '( 4417-9 ) ( 4417-9 2 5 p. Fe -3 p. Ti 4434-6 4434-0 1 4437-7 1 4437-6 3 4434 4 1 Ast. Unk. 4470-1 4469-9 1 4471-6 6 4471-6 8 4471-6 1 4171-6 1 -2 He 4480-7 , 1474-9 4481-6) ( 4487-2' 20 4481-3 7 4481-3 4 4481-3 6 81 , 8 p. Mg. 4509-9 4511-2 2 4508 5 1 4508-5 1-2 4508 5 5 p. Fe 4529-5 4528 1 2 4529-6 1 4529-6 1 - p.Ti 4549-2 4549 -8 3 [ p. Fe \ P- Ti * .' O in Bellatrix. 10 SOLAR PHYSICS COMMITTEE. COMPARISON OF THE "DARK LINE" SPECTRUM OF NOVA PERSEI, &c. continued. Nova Persei. Pickering's Absorption. Spectrum, Feb. 22, 1901. Riga S. Kensington. Bellatrix. S. Kensington. Markab, S. Kensington. a Cygni, S. Kensington. Origins in Type- A A Negative (1] 4552 -4 A A Negative (2) Int Max. x A = 40. 0) A. A. fnt. (10) A A Int. (10) A A Stars. Int. (10) 2 4552-8 1 4552-8 3 4552-8 <1 p. Si (III.) 4568-0 <1 4568-0 2 4568-0 <1 p. Si (II I.) 4571-9 1 4570-4 1 4570-6 <1 4572-2 1 4572-2 4 p. Ti 4615-4 1 4616-8 2 p. Cr 4642-7 4642.2 2 4642-5 1 4640-1 1 4641-4 <1 4665-0 4661-1 4 ; 4665-5 <1 4713-7 4714-4 3 4713-3 ; 3 4713-3 4 He 4739-5 ] ! 4861-5 C 4854 -7, '4861-5 (.4867V 40 4861-5 10 4861-5 10 4861-5 10 4861-5 10 H,, 4885-2 9 4884-3 1-2 p. Cr 4913-7 i ' 4922-2 2 4922-1 5 4922-2 5 Ast. 4924 -1 8 p. Fe 5019-4 1 5018-6 5 5015-7 3 Ast. p. Fe 5325-2 1 5399-2 5393-5 1 5431-9 5429-2 1 . - - - 5455 ' 3 1 . 5634-8 1 5663-2 1 5677 1 2 5695 2 5693 3 6 . 5719-3 5717-6 5 5761-2 5758-0 1 _ 4128-1 (7). A Si. line (p. Si II. 4128 '1) is near this position, but as the companion 4131'! is not in the nova spectrum the p. Si origin must he abandoned. PHENOMENA OF NEW STARS. II 4151 '5(1). In Bellatrix 4150 '5 is ascribed possibly to X, but the other X* lines are present, whereas in the nova spectrum they are not. Again the likeness to the type-star breaks down. 4266'4(2). In Rigel and Bellatrix the line at X 4267 '5 is p.C. As (4207-5) Asterium and Helium, elements usually associated with Carbon I., are present, there is nothing l^eyond the divergence of wave-length and possibly the absence of X' to dismiss p.C as a possible origin. In the nova it is weak (2 on a scale of 40), and is therefore more like Bellatrix or Alnitam than Rigel. 4366 '4 or 4365 '5 (1). The only representative in the type-stars is a very weak line, ascribed to p.Mn, in a Cygni. 4388'! (3). In Rigel and Bellatrix this is ascribed to Asterium, and at first sight one would accept this origin for the nova line. But 3964 '9, 4009 '4, 4143 "9, as strong, or stronger, in the other stars, as well as other, weaker lines, are not shown in the nova spectrum. Of the 11 Asterium lines occurring in Bellatrix, only two, 4388 and 4922, are possibly represented in the nova spectrum. There are only three Asterium lines possibly represented in the S. Kensington spectrum of a Cygni, and one of these is 4388'!, but 4922 is not in a Cygni, while the other two Asterium lines in the spectrum of that star do not appear in the nova spectrum. We must conclude that the presence of Asterium in the nova is open to grave doubt. '- . In Bellatrix these weak nova lines are ascribed to oxygen and 44o4 ' b ! 1 ) ) nitrogen respectively, but in the nova many other oxygen and nitrogen lines are absent, and the oxygen and nitrogen origins are not tenable. [ In Rigel the intensity of this line is 6 on the S. Kensington 4471 ' 6 scale of 10, whereas in the nova Pickering gives the intensity as 1 on a scale of 40. The line 4121 is not present in the nova, but all the other He lines present in Bellatrix are. In Bellatrix the intensity of 4471 on the 10 scale is 8. It would appear, then, that helium is probably represented in this nova absorption spectrum, but certainly not in the same condition as in Rigel or Bellatrix. The relation in intensity to the Mg line at 4481 is certainly not like Rigel or Bellatrix, it is more like Markab, but even there there is a marked difference. In Sirius 4471 (? helium) is < 1 and 4481 is 7 (on a scale of 10), and this gives a better parallel to Pickering's relative intensities. SOLAR PHYSICS COMMITTEE. '0 (20). In the type-stars this line is due to p. Mg, hut it is always a hard, sharp line. In the nova 41-81 '0 is the centre of a dark b a nd (c.tl.) which extends from 1471 "9 to 4487 '2, a breadth of 12 '3 A. In intensity it is only exceeded by the hydrogen lines. One concludes that p. Mg is probably present, but under conditions differing considerably from those obtaining in the ordinary helium stars. The other p. Mg line 4395 is not present. * 4529 '5 (2). An unknown line ocurring in 7 Orionis (i) and in a Cygni (1) and Sirius. In the latter two it may be due to p. Ti at 4529 '6. 4552 '4 (2). In Bellatrix this is an intensity 3 line, probably due to p. Si (III.) 4552 '8, in a Cygni (< 1), in lligel (1). But in each of these cases the other neighbouring p. Si (III.) lines are also present, whereas in the nova they are not. Hence the silicon origin is untenable for the nova line. 4571 '9 (1). In Bellatrix there is a faint unknown line (1) at X 4570 '4, which also possibly occurs as X 4570 '6 in a Cygni (< 1). In Procyon there is a line at X 4^70 -i \ (6), probably due to Mg-p Ti, but in the nova these origins are not otherwise represented, so we cannot accept this likeness to Procyon ; this also applies as regards the near line X 4572*2 in a Pegasi. 4615 '4 (1). No such line in the type-stars and no origin suggests itself. 4642-2 (2). An unknown line in Rigel at X 4642 '5 (1). 4665-0) 4661 ' known lines of these wave-lengths in the type-stars. ' \ is the helium line found in the type stars, but its intensity 4/13 '3 } in the nova spectrum, compared with the lines 4471 and 4481, is abnormal. XX Intensities in Nova. Max. = 40. Rigel. = 10. Bellntrix. = 10. Markab. = 10. a Cv^ni. ='io. 4471 1 6 8 1 1-2 4481 20 7 4 6 8 4713 3 3 4 4922 ' 2 (2) in Bellatrix is probably due to Asterium (int. 5) while in a Cygni there is a line (int. ==8) at 4924 '1, due to p. Ee. But the other lines due to Asterium (see note on 4388'!) are absent from the nova, and many p. Fe lines are lacking. PHENOMENA OF NEW STARS. 13 A study of this table, then, leads to the conclusion that the dark-line spectrum of Nova Persei, stated by Pickering to be like Rigel, is like no spectrum we meet with in typical stars. Pickering's wave-lengths are too uncertain to permit any rigid conclusions being based on them, but, giving them the broadest interpretation possible, we can only conclude that the spectrum is not Kigelian in character. The hydrogen lines are common to both, and it may be supposed, although there are several difficulties, that helium, asterium, and magnesium are also present. If we consider these elements, we see that the spectrum is far more like a Cygni than Rigel. The helium line 4026 '4 is much too weak for it to be Rigeliau or Crucian, and so are the lines at XX 4388 and 4471. This last gives us a very striking comparison with the adjacent line, 4481 Mg. X / Inteniities in Nova. Max. = 40. Kigelian. = 10. Crucian. = 10. Markabian. = 10. Cygnian. = 10. He. 4471 1 6 8 1 1-2 Mg. 4481 20 7 4 6 8 This discussion leads us to the conclusion that the evidence is certainly not in favour of the spectrum being considered as Rigelian. In fact it is almost impossible to tie it up to the spectrum of any formed star of the normal type, for whichever line we take we find several important inconsistencies. In the second case of a nova spectrum showing dark lines only, in its first stage, Miss Cannon describes the spectrum of Nova Geminorum (2) on March 13, 1912, as " totally unlike that of any other nova so far observed, and resembled " that of the sun, except that the hydrogen lines were stronger. It appeared, " therefore, to belong to Class F5" ::: ", having also several of the dark lines " unusually well marked."! As Miss Cannon points out, this is only the second case in which the spectrum has been photographed while the nova was increasing in brightness to its primary maximum, and in each case a strong continuous spectrum crossed by dark lines, as in a normal star, has been recorded. As we have already shown, it is difficult to establish any detailed similarity between the dark-line spectrum of Nova Persei (2), recorded by Pickering and that of any normal star, but in the case of Nova Geminorum (2) the resemblance of the spectrum to thai of Procyon is very striking. Copies of the two spectra, taken with the same dispersion and under similar conditions, are reproduced on * Procyonian. t Harv. Coll. Obs. Circular No. 176, p. 2. a 19881 14 SOLAR PHYSICS COMMITTEE. a plate accompanying the Harvard " Circular ", and they present no differences that are seen at first glance. More careful examination arouses the suspicion that the spectrum is slightly stronger on the less refrangible edges of H. K., Kb, and H 7 , in the nova spectrum than in that of Procyon, hut this apparently increased intensity is not to he compared with the intensities of the bright bands that subsequently appear in these positions. There is also a slight difference of the grouping of the lines between H<5 and H 7 , the grouping in the nova spectrum being more like that in the later nova spectra (of the bright- line stage), than that in the spectrum of Procyon. Miss Cannon directs attention to this feature in saying that lines near X 4175 are stronger in the nova spectrum, and it may be noted that there is a strong pair of enhanced iron ( 4173-52) . lines -j / , 1 HQ.f K ( m this re S lon which are conspicuous in a Cygm and come out (, 41 /o ' 95 ) strongly, as will be seen later, in the subsequent stage of novse. The spectra reproduced by Prof. Pickering also show a stronger continuous spectrum in the ultra-violet in the case of the nova than in that of Procyon. The conclusion to be drawn then, in the case of Nova Geminorum No. 2, is that a dark-line spectrum, very similar to that of Procyon, but possessing features suggestive of the later nova stage, existed during the short time that the star was rising to its primary maximum. 3 ( but there is no proljability of its identity with this pair, for, as has been shown above, ' ' 4640 " only Incomes a strong band as the enhanced bines disappear from the spectrum. From the general, frequently very broad and hazy, appearance of the line in most cases we should expect the origin to be gaseous, but no known gas under known conditions gives an isolated radiation in this position. Another conclusion that appears well founded is that either the chemical nature of the origin, or the condition under which this radiation can appear, is very unstable, the whole possible existence lying sharply defined between the enhanced metallic line and the nebular condition. This peculiar condition, like the enhanced line condition, never lasts very long. Whether the changes producing these conditions are analogous to a rise or a fall in temperature, it is important to note them, for they afford evidence of changes taking place in small masses. The X 4640 radiation has no counterpart among the absorption lines of normal metallic stars, nor has it been found in meteorites examined terrestrially; but, as is shown in Table VI., it has been found in several nebulae, especially those of a planetary nature. Tliis is to be expected, for in the nova the radiation is from a body about to assume a nebular character. The fact also points to those special nebulae in which the radiation is found being in a different, probably a transitional, stage from those not showing it, and it may be worth while to look for evidence of change in such bodies. Apparently the X 4640 radiation is not found in the spectrum of any ordinary irregular nebulae, e.g., that of Orion, thus differentiating them from the planetary nebulae. Some of the light sources in which the line has been observed are set out in Table VI. 26 SOLAR PHYSICS COMMITTEE. TABLE VI. KOVJE, NEBULAE, &c., IN WHICH THE BRIGHT BAND AT X 4640 HAS BEEN RECORDED. Object. A. given. Date. Remarks and Reference. N. Auriga? 4630 Aug. 1892 Int. = 7 (max. = 100). Campbell (Astronomy and Astrophysics, XII., p. 726, 1893). N. CarinsB About 4700 Apr. 14, 1895 As brigbt as the hydrogen lines. Pickering (Astro- physical Journal, II., p. 320). N. Sagittarii (1) 4643 Apr. 19, 1898 Bright band. Pickering (Astrophysical Journal, IX., p. 183). > 462 Apr. 21, 1898 Int. 4 (where H /5 = 10, X 496 = 20, and X 501 = 60). Campbell (ibid., p. 308). N. Persei (2) 464 Apr. 12, 1901 Strongest line in spectrum. South Kensington. N. Geminorum (1) 4647 Mar. 28, 1903 Strongest line in spectrum. Hale (Astrophysical Journal, XVII., p. 305). " > . 4625 Apr. 22, 1903 Predominant line. Reese and Curtis (Astrophysical Journal, XVIII., p. 300). N. Aquilse (2) - 4646 Aug. 18, 1905 Brighter than X 4472 ; weaker than H lines. Picker- ing (H.C.O. Circular 106). N. Velorum 4640 H.C.O. Circular, 131. N. Geminorum (2) 4639-2 Apr. 15, 1912 Strong band, probably double. South Kensington. P. Cygni - 4631 Frost identifies with nitrogen. (Aslrophysical Journal, Vol. XXXV., p. 2H7). Planetary nebula, 463 Bright line ; nebular lines very bright. Campbell N.G.C. 7,027. (Astronomy and Astrophysics, Vol. XIII., p. 498). Planetary nebula in 464 Very faint bright line ; nebular lines verv bright. Draco, N.G.C. 6,543. Campbell (ibid.). Planetary nebula, 4643 Faint bright line. Campbell (ibid.). G.C. 4,390. Planetary nebula, 4645 Faint bright line. Campbell (ibid.). N.G.C. 7,662. Wolf-Rayet Stars 4636 Weak, bright ; appears in 10 stars of 32 discussed generally. by Campbell. (Astronomy and Astrophysics. Vol. XIII., p. 470.) Type-Stars of Miss Cannon's Oa., Oc., Od., and Oe. classes. Oa.-Cariiiffi A. G.C. 4633 Most conspicuous band, bright-line star. . Puppis 15,305. hydrogen lines faint. (H.C.O. Annals, XXVIII., Pt. 2, p. 146.) Oc.-Scorpii, A. G.C. 4633 Bright band star. f. Puppis hydrogen present, 22,763. probably 4471 He. 4688 most conspicuous. (H.C.O. Annals, XXVIII., Pt. 2, p. 148.) Od. '(,. Puppis 4633 Fairly strong, bright (with 4688) in an otherwise dark-line spectrum containing H and p. II lines. Pickering. (H.C.O. Annals, XXVIII., Pt, 2, p. 148). Oe. 29 Canis Maj. 4633 This line and X 4688 the only bright lines in a spectrum of many dark lines. (H.C.O. Annals, XXVIII., Pt. 2, pp. 146-148.) PHENOMENA OF NEW STARS. 27 3 (A). THE "NEBULA" SPECTRUM OF XOV.E. From the 4640 stage the spectra of novae generally change gradually to one in which the ordinary "nebula" spectrum is the dominant feature. This change proceeds until the X 500 line and other lines in the ultra-violet characteristic of nebulae are the strongest in the spectrum. The nebular spectrum observed in the various novae is given in Table VII. with Wright's* general spectrum of nebulae for comparison. From this table it appears to be beyond doubt that in the fourth stage of the development of the spectra of new stars the radiations emitted are the same as those emitted by nebulae in general ; the minor differences are not so important as they frequently are in passing from one nebula to another. TABLE VII. SPECTRA IN THE "NEBULA" STAGE. Nebular Lines (Wright). AA Novas Remarks. Aurigse (Campbell), Aug. 1892. AA Persei (S. Kensington), 26 Aug. 1901. AA Geminorum (1) (Perrine), Apr. 1903. AA Lacertae (Wright), 30 Mar. 1911. AA Geminorum (2) (Cambridge). 13 Aug. 1912. AA 3726-4 342 : Probably the line independently re- corded by Dr. vou Gothard. 3729-0 3835-8 384 H, 3868-88 3868 3889-14 389 11' 3965 1 396 3967-65 3970-23 3970 397 H. 4026-7 4068-8 4101-91 4098 4102 410 t*A 423 426 4340-62 4336 4341 434 H y * Astrophysical Journal, Vol. XV*!., p. 53, 1902. 28 SOLAR PHYSICS COMMITTEE. NOV^E SPECTRA IN THE "NEBULA" STAGE continued. Nebular Lines (Wright). AA Novas Remarks. Aurigae (Campbell), Aug. 1892. AA Persei (S. Kensington), 26 Aue. 1901. AA Geminorum (1) (Perrine), Apr. 1903. AA Lacertae (Wright), 30 Mar. 1911. AA Geminorum (2) (Cambridge), 13 Aug. 1912. AA 4363-37 4358 4364 . 4471-71 4466 4471 446 451 460 ' 4630 4636 4625 464 4685 73 4681 4684 471 ' 4720 4740 4808 4861-54 4857 4862 486 4861 486 % 4959-05 4953 4959 4959 496 2nd nebular line. 5006-89 5002 5007 5007 501 1st nebular line. 531 5752 575 6583 3 0'). ON THE SPECTRUM- OF NOVA GEMINORUM (2). PRELIMINARY GENERAL DESCRIPTION. Owing to the adverse weather conditions, no successful photographs of the spectrum of Nova Geminorum (2), during its "bright-line" stage, were secured at South Kensington. Ey the kindness of Prof. Iniguez of the Madrid Observatory I have had placed at my disposal, for analysis and discussion, copies (positives on glass) of the admirable photographs of the spectra obtained at Madrid. The dates of these extend from March 16 to April 16 inclusive, during which period the spectrum was photographed on 26 days. With the exception of March 26, when, owing to bad weather, the photograph was not very satisfactory, all these have PHENOMENA OF NEW STARS. 29 been copied at South Kensington by direct contact, and enlarged prints, four and a half times the originals, made from the resulting negatives. These enlarge- ments have been placed in a sequential series and are reproduced in Plate II. To facilitate the interpretation of the spectra, several stellar spectra in the Kensington series, such as a, Cygni and e Orionis, have been copied to the same scale as the Madrid spectra. Without going into much detail here, it may be said that a comparison with a Cygni showed at once that the chief bright lines in the earlier nova spectra agree in position with conspicuous lines in a Cygni, and there remains little doubt that the two sets of lines are identical in origin. With some modifications the chief lines in the Nova Geminorum spectra agree also with those found in previous novae, e.g., Nova Aurigae and Nova Persei (2). From March 24 these a Cygni lines decline in intensity and other lines develop, chief of which is a strong, isolated, sometimes double, line at X 4639, seen most conspicuously in the spectra of April 6 and 15, The earlier photographs show another customary feature of novae spectra in that the bright hydrogen lines are accompanied on the more refrangible side by well-marked dark lines. These vary in structure from day to day. On some days they are beautifully divided centrally by a sharp radiation line or reversal. In other cases the dividing line is broader and the two dark component parts of the whole absorption line are very unequal in intensity. These variations are emphasised in the three successive photographs of March 24, 25 and 27. Taking H Y as an example, on the first date the absorption line is divided by a comparatively narrow bright line and the two portions of the dark line are fairly equal in intensity, the more refrangible being slightly the stronger. On the next day the dividing bright line is a little broader and the difference in intensity between the two dark components is accentuated. On March 27 the dividing bright line is still broader and the stronger dark component is now on the less refrangible side. There is distinct evidence, also, in the photographs of March 24 and 25, that not only are the hydrogen lines accompanied by dark lines, but the bright line in the position of the helium line 4471 and the bright enhanced iron lines at XX. 4176, 4233 have similar accompanying broad dark lines, these also being divided by a narrow bright line, but not in such a decisive manner as in the case of some of the hydrogen lines. The detailed reduction and discussion of these spectra, carried out by Mr. Baxandall, is given below. In Table VIII. appear the wave-lengths and probable origins of the bright bands seen on the negative of March 19 when the enhanced lines of a Cygni were predominant. a 19881 E SOLAR PHYSICS COMMITTEE. 3934-6 3960-2 } 3969-2 ( 3976-3 ( 3979-2 ) 4006-9 4019-3 4061-0 4081-5 4090-7 '4101-85 4109-6 4112-3 4154-4 '- ) o 4161 4174 4182-0 4188-0 4222-1 4232-5 4242-9 I 4252-8 4267-2 429T3 4301-1 4311-1 4317-3 4328-5 *4340-66 } 4353-3 I 4357-9 4376-4 4392-3 4408-7 4423-8 TABLE VIII. MADRID SPECTRUM, MARCH 19. * = Fiducial lines used for reduction to wave-lengths. H Probable A of Probable Origin. Origin. Remarks. Ca 3933-83 Middle of K (Ca). ( Violet edge of H. II 3968-63 J Middle of broad H (+ H Ca). Ca 3970-25 1 Middle of strong maximum in Hj. i ( Red edge of H . Weak narrow radiation. Middle of broad weak line with ill-defined edges. II ? Reversal in dark H<5. j Violet edge of Us. H 4101-85 Middle of Us. Middle of strong maximum in Hj. ( Red edge of Us. Detached narrow bit of radiation. p. Fe j 4173-61 | , 17fi .a,j 4179-05 f 4176 3o j Violet edge of broad line. Middle of strong broad line. Red edge of strong broad line. Violet edge of broad bright line. p. Fe 4233 32 Middle of broad bright line. ,' Red edge of broad bright line. Weak narrow radiation. 4340-66 Violet edge of broad bright line. Middle of broad bright line. Bed edge of broad bright line. Detached narrow radiation on violet side of 11-y, proba- bly reversal in dark H-y. Violet edge of H-y. Middle of H 7 line. Red edge of H-y. Narrow radiation bordering H-y. Probably the less refrangible portion of a broad line, part of which is hidden by bright H-y. tl>is broad line being probably identical with the a Cygni proto-iron line 4351-9. t Narrow strong radiation. Narrow and rather weak radiation. Narrow and fairly strong radiation. PHENOMENA OF NEW STARS. 31 MADRID SPECTRUM, MARCH 19 continued. A Probable A of Probable Origin. Origin. Remarks. 4434-2 Detached narrow bit of radiation. 4450-4 >, , 4456-4 ) 4470-0 i 4483-3 ) He 4471-65 Violet edge of weak broad bright line. Middle of weak broad bright line. Red edge cf weak broad bright line. 4496-8 ^ 4498-5 4514-7 } 4516-4 4532-8 J p.Fe j 4508-46] 4515-51 f Mean 4520-40 (4516-78 ) 4522-77 ) Extreme violet edge of broad line. Detached bit of radiation in broad line. Position of maximum intensity in broad line. Middle of hroad line. Red edge of broad line. s 4537-4 ) 4549-0 4560-2 ) p. Fe 4549-64 Violet edge of broad line. Middle of broad line. Red edge of hroad line. 4572-7 ) 4584-1 [ 4594-9 ) p. Fe 4584-02 Violet edge of broad line. Middle of broad line. Red edge of broad line. 4623-4 ) 4633-1 4643-9 ) p.Fe j 4629-51 I 4639 . 50 4635-50 [ 4 Violet edge of broad line. Middle of broad line. Red edge of broad line. 4660-2 Middle of broad line, the violet edge of which merges into 4633 line. 4674-7 Diffuse red edge of broad line. 4698-1 ' Weak radiation. 4727-8 .; <> 4848-1 ) ! *4861- 49 i H 4875-0 ) 4861-49 /Violet edge of H/8. Middle of H/s line. Red edge of H0. For the type of spectrum where the enhanced lines have nearly disappeared and of which the chief feature is the hroad and apparently double line near X 4640, the spectrum of April 7 has heen measured. The results are given in the following Table IX. The width of the whole band at A 4640, taken in its dual aspect, is about 70 tenth-metres. TABLE IX. MADRID SPECTRUM, APRIL 7. * = Fiducial lines used for reduction to wave-lengths. A Probable Origin. A of Probable Origin. Remarks. 3889-1 H 3889-1 Middle of H^ line. 3970-2 H 3970-25 Middle of H, line. "4101-85 H 4101-85 Middle of bright Hj line. 4174-5 p. Fe j 4173-61 \,.^ K , I 4179-05 i 417633 Middle of weak broad line. SOLAR PHYSICS COMMITLEE. MADRID SPECTRUM, APRIL 7 continued. A Probable Origin. A. of Probable Origin. i Remarks. 4232 5 p. Fe 4233-32 Middle of weak broad line. 4263-5 4301 -1 p. Fe ( 4296-72 |S-34 430 - 3 Middle of weak broad line. A 330-0 *4340-66 4351-0 H 4340-66 j Violet edge of bright H y . I Middle of bright H y . ( Red edge of bright H Y . 4470 '3 He 4471-65 Middle of weak broad line. ( 4508-46 4516-4 x p. Fe J 4515-51 Mean j 4520-40 4516-75 (4522-77 Middle of rather weak broad line. 4549-0 p. Fe 4549-64 Middle of weak broad line. 4584-1 p. Fe 4584-02 Middle of rather weak broad line. 4602-4 i J 628-6 4644-7 f Weak violet edge of broad bright band. Middle of brightest portion of broad bright band. Middle of dividing dark line in compound broad bright <; band. 4059-6 4672-2 J Middle of weak less refrangible portion of broad bright band. . Red edge of broad bright band. 4861-49 H 4861-49 Middle of bright H,,. For the type of spectrum which shows the line near X 4639 as strong and isolated, the photograph of April 15th has been selected for measurement. The details are given in Table X. The position of the line near 4639 was independently measured by three observers, and the wave-length given is the mean of the three measures. TABLE X. MADRID SPECTRUM, APRIL 15. * = Fiducial lines used for reduction to wave-lengths. \ Probable Origin. , Xof Kemarks. Probable Origin. 3970-25 H 3970-25 Middle of bright H,. 4101-85 H 4101-85 Middle of bright IJj. *4340-66 H 4340-66 Middle of bright H r 4470-2 He 4471-65 Middle of weak broad line. 4639-2 Middle of strong isolated bright line. 4680-2 Middle of weak broad line. '4861-49 H 4861-49 Middle of bright H^ NOTE. In addition to the lines in the above list there are three weak lines between H and Hj ; three between Hj and H y and five between H y and Hp These are so weak, however, as to be almost lost when the magnification of the measuring machine is used on them. Those between H s and H y are probably the lines XX 4176, 4233, and 4264. PHENOMENA OF NEW STARS. 33 DETAILED DISCUSSION OF THE SPECTRUM OE NOVA GEMINORUM (2), MARCH 16 APRIL 16, 1912. In addition to the main features of the spectrum of Nova Geminorum (2), already discussed in the preliminary general statement, and shown in Tables VIII., IX., and X., there were many less conspicuous but important details shown in the plates ; also very striking changes took place throughout the spectrum from time to time, such as were recorded for the spectra of the earlier novae. These details and changes are dealt with by Mr. Baxandall in the discussion that follows, the spectrum for each day being treated separately. See Plate II. March 16. The bright hydrogen lines H# H 7 H^ H t are shown, H 7 being the strongest. Lines H& H 7 , and HJ appear to have a fine reversal down the centre. This is best seen in the H 7 line. On the more refrangible side of H 7 is a well-defined absorption line, the width of which is about one-third that of the bright H 7 line. There is also a suspicion of absorption lines to the left of Hjs and H, but these are not so definite as that at H 7 . The K line of calcium and probably the H line (superposed on H f ) are seen as rather faint broad radiations. Between H 7 and Hn the usual nova bands at XX 4176, 1233, 4300 are seen, but not conspicuously. These are to be ascribed to the enhanced lines of iron at (4173-61 , (4296 '72 , . , 1 4179" 0" 42->3 32 anc ^ j i AQQ Q f which are so well shown in a Cygni. Between the last of these and the dark H 7 , and fringing the latter, there is another narrow radiation, which, however, may be a portion of the complex structure of the whole H 7 radiation. Assuming the apparent middle of the H 7 absorption line to be the real middle, which may easily be a false assumption, this is displaced from the fine dark line in the middle of the broad bright H 7 an amount equivalent to 15 "3 tenth-metres, which, if the displacement is assumed to be due to a velocity effect (another assumption Avhich may be proved false by future research), shows a relative velocity of rather more than 1,000 kins, per second for the two bodies producing the dark and bright hydrogen lines. Prom H 7 to about X 4650 there is a complex series of faint and rather broad radiations with apparent fine absorption lines amongst them. Judging from the de- velopment of these broad lines or bands in the nova spectra of subsequent dates it is fairly certain that most of the faint bright bands in the spectrum of March 16 can be ascribed to enhanced lines of iron which occur prominently in a Cygni. Without committing oneself definitely to the assertion that the. apparent dark lines are only, in general, interspaces between broad bright bands, it appears likely that such is the case. March 17. The hydrogen lines H/s H 7 H^ H (superposed on H calcium) and the K line of calcium have developed considerably in intensity, the first three showing rather complex structure. There are two, possibly three, distinct maxima 34 SOLAR PHYSICS COMMITTEE. of brightness, the least refrangible being the brightest. The difficulty of deciding whether the most refrangible of the three maxima in H 7 and H^ is really part of the hydrogen radiation rests in the fact that the Madrid spectra are taken with an objective prism and there is no laboratory comparison spectrum to enable us to say which are really the middles of the complex hydrogen lines. The H e and K radiations are rather diffuse and show little structure. The enhanced- iron radiations at XX 4176, 4233, 4300 have strengthened from the previous day, and the same remark applies to the complex series of bands between H 7 and X 4650. The identification of most of these individual lines with laboratory lines is made very difficult owing to a succession of probably overlapping images, but the fact that in other parts of the nova spectra Avhere the enhanced lines of iron are more isolated (between H, and He and on the less refrangible side of H0) all these special lines of the element named agree in position with well- marked nova lines, taken together with the fact that there are about seventeen enhanced iron lines in the region H 7 to 4650, where the nova spectrum is so complex, would appear to' make it almost certain that the enhanced lines of iron are the chief factor in the nova spectrum. Another significant fact is that in the region X 4650 to H0, where no enhanced lines of iron occur, the nova spectrum is almost a complete blank. On the less refrangible side of H in the spectrum of March 17th there is what appears to be a broad absorption line, but whether this is a genuine absorption line or only an interspace between the bright H 7 line and the next bright line on the less refrangible side of H 7 is doubtful. The structure of the 4300 line (probably a double in itself) combined with the complex structure of the neighbouring H 7 line gives the appearance of a set of six fairly sharp emission lines with dark interspaces or absorption lines between. As, however, it is fairly evident that the nova spectrum consists chiefly of broad bright lines showing considerable structure in themselves, it seems almost useless to measure the position of each bit of radiation and each bit of apparent absorption, and use the resulting wave-length as a means of possibly identifying these bits with lines of chemical elements. Such a method is almost certain to give fictitious and misleading identifications. In dealing with -novae spectra it must be borne in mind that the number of radiation lines is probably small, and that the type of line is broad and one showing some structure. In the region of H 7 to X 4650 the probably overlapping images of the individual lines makes it difficult, if not impossible, to tell which are the real middles of the lines. A similar difficulty presents itself in deciding whether the narrow dark portions of the spectrum in this region are real absorption lines or interspaces between bright lines or bands. Judging from their narrowness, as compared with the width of the well-authenticated bright lines, the latter is more likely to be the real interpretation. March 18. There i not much difference between the spectrum of this date and that of the 17th. The structure of the hydrogen lines shows the less PHENOMENA OF NEW STARS. 35 refrangible portion of the bright lines distinctly stronger than the remaining portion. The extra radiation to the left of H 7 and IU, which on the 17th gave the appearance of a possible third maximum in the radiations, is entirely missing. H and K are about as on the 17th. The three Ixmds XX 1176, 4233, 4300 are the only ones well visible between Hs and H 7 . The dark hydrogen lines to the violet side of the bright ones are present but not conspicuous. March 19. The hydrogen lines are now very strong, H> and H y being divided in the middle by an apparent absorption line. The two bright components in the structure of these lines are now nearly equal in intensity. The bright He line does not show the same structure as the others. It is distinctly brighter at the red edge and then keeps fairly uniform in intensity up to the violet edge, giving the appearance of a winged line rather than a double as at H-,, and H0. The H ( line is of a dual nature, the less refrangible component being much stronger and better defined than the more refrangible. The K line is also double, but the difference in intensity of the two components is not nearly so well marked as in H which probably involves the H calcium line. The broad lines XX 4176, 4233, 4300, are now quite conspicuous and have more definite edges than in previous photographs. They also show somewhat similar structure to the hydrogen lines H r and H0. The lines between H 7 and X 4650 show tetter definition. Several of these now show the same structure as the three lines between H 7 and He ; being broad lines with fairly well developed edges and divided down their width by fine dark lines. There can be little doubt f 4508 '46^ that the enhanced lines of iron at XX ~9A.j,or> 4."fi-ftQf' ^^4 '02 and NgoK-KAp & 11 we ll represented in a Cygni, are involved in the four bright l>ands in this region. Which these lines are will be gathered from reference to Plates I. and II. Owing, probably to the varied bunching of these sets of lines, the structure in the different nova lines is not quite the same. The apparent absorption lines between the members of this set of bright broad lines have probably no significance as spectrum lines, being only interspaces. The lines H0 Hy and H<- are accompanied by narrow dark lines not very conspicuous, especially in the reproduction in Plate II. March 20. Very similar to spectrum of March 19, except that there is not so much detail shown in the individual lines. The brighter portion of the H. line is not so accentuated here as on the 19th. Two or three weak diffuse lines have developed in the region between 4650 and H,,, which region was in previous spectra almost void of radiation lines. The enhanced iron line 4924 '1, conspicuous in a Cygni and kindred stars, is fairly well shown. Lines H^, H Y , and H 4 are still bordered by rather broad absorption lines, but of no great density. At H y and W t these absorption lines are divided in the middle by a weak radiation. If the H0 absorption line is similarly divided it is not very evident. 36 SOLAR PHYSICS COMMITTEE. March 22. Of the spectra up to and including this date, that of March 22 shows most detail in the spectrum lines. The most conspicuous feature is undouhtedly the very black absorption lines which accompany the bright hydrogen lines, especially H^, H y) and H a . These are about three-quarters the width of the corresponding bright lines, and H y and H 5 are beautifully divided in the centre by sharp radiation lines or reversals. There is a suspicion that dark H^ is similai'ly divided. The bright hydrogen lines are still not uniform in intensity throughout their width, there being two maxima of brightness fairly equally divided by a portion of considerably less intensity. Between H/s and H the out- standing lines are still those at XX 4176, 4233, and 4300, which show somewhat similar structure to the hydrogen lines and are about the same width. Line X 4300, however, seems more split up than the other two, and shows two maxima of intensity more conspicuously than they do. The intensity curve of this line is not symmetrical. On the violet side of lines XX 4176 and 4233 there are broad absorption lines similar to the dark hydrogen lines, but not nearly so dense. These absorption lines also show similar structure to the dark hydrogen lines, being divided down their centres, but not so definitely as in the hydrogen lines. A few minor radiations have here developed in addition to the three main bands. The spectrum between H v and H p still shows a complex system of radiation lines or bands of varying width, but in most cases they can be called, broad lines. Between these are what appear to be narrow absorption lines, but the probability is that these are only gaps between bright lines, and their estimated wave-lengths probably have no significance. At any rate, if these are assumed to be genuine lines, to be consistent one should estimate the positions of each portion of the dark hydrogen lines, and adopt these as wave-lengths of absorption lines. It is scarcely necessary to point out that such a procedure would give wave-length numbers of no value in so far as identifying the lines of chemical elements is concerned. Prom the wave-length values obtained from the middles of the broad radiation lines it is fairly evident that many of these agree with strong lines in the spectrum of a Cygni, which in the majority of cases have been shown to be identical with enhanced lines of iron. On the less refrangible side of H^ the enhanced iron lines XX 4924 '11 and 5018 '63 (very conspicuous in a Cygni) are easily seen in the spectrum of this date. March 23. The spectrum of this date is practically a replica of that of the 22nd ; the only striking difference is that the radiation band, the middle of which agrees with the position of the X 4471 line of helium, is now somewhat narrower, and the absorption line or interspace on the violet side of it correspondingly broader than on the 22nd. March 24. Very similar to the spectra of the 22hd and 23rd. The only notable differences from the spectrum of the 23rd are: (1) The absorption lines bordering the bright hydrogen lines are denser, and thus show more contrast from the bright lines. PHENOMENA OF NEW STARS. 37 (2) The absorption lines on the violet side of the nova bands at XX 4176 and 4233 are more conspicuous and show, more distinctly than heretofore, the same structure as the dark hydrogen lines, viz., they are divided down the centre by a radiation line or reversal. (3) The absorption line to the left of the bright line at X 4471 shows more definite structure than previously. It is divided centrally by a fairly fine bright line, but the less refrangible portion of the absorption line taken as a whole is far blacker and more conspicuous than the more refrangible portion. (4) The broad radiation band, the middle of which is at X 4549, shows a little more structure, as it is now crossed by an apparently thin absorption line at about one-third of its width inside the less refrangible edge. All the lines referred to in the spectrum of this date will easily be recognised by reference to Plates I. and II. March 25. The spectrum shows some very striking changes from those of the three previous days, which were much alike. There appears to be far more continuous spectrum than in any of the earlier spectra, and the radiation lines or bands, although they are identical in position with those in the other spectra, do not stand out prominently. The most striking change, however, is in the dark hydrogen lines. In the spectra of March 22, 23, and 24 they were very conspicuous, broad, and dense lines, with fine bright dividing lines down the centres. On the 25th the dividing line is considerably wider and more noticeable, and the more refrangible portions of the dark hydrogen line is in each case stronger than the portion bordering the broad bright hydrogen lines. In fact, what is probably only a portion of the width of the whole absorption line of hydrogen stands out quite prominently as an apparent narrow and sharp absorption line. The bright hydrogen lines have a very sharp edge on the violet side, but are very diffuse on the red side, giving the appearance almost of a fluting. The structure in the bright lines is not very apparent. March 27. The bright hydrogen lines are still very conspicuous. The dark hydrogen lines have again changed considerably. The most refrangible part which stood out so prominently on March 25 is now very weak, the less refrangible portion tordering the strong bright hydrogen lines being the more conspicuous. The bright hydrogen lines now show a broad and intense portion very sharply cut on the more refrangible side, less so on the less refrangible side, with some structure shown in the middle. There is another part of the hydrogen radiation not very broad and far less intense than the main portion on the red side of the. latter and giving rather the appearance, on the whole, of a strong bright line winged on the red side. There is in the Madrid positive a weak absorption line dividing the very bright portion of the hydrogen line from the less bright portion, but this is not so evident in the reproduction. a 19881 F 38 SOLAR PHYSICS COMMITTEE. The bands at XX 4176, 4233, and 4300 are present but faint. In the spectrum between H 7 and H# striking changes have taken place from the 25th. Prom H 7 to about X 4500 most of the lines seen in previous spectra have now practically disappeared, leaving as the only outstanding feature between H 7 and H^ the three bands at XX 4515, 4549, and 4584, all present in previous spectra, and a much stronger broad band (probably a double broad line), whose mean position is about X 4640 and width about 60 tenth-metres. The broad band is probably not a development of the pair of broad lines occupying an approximately similar position in the earlier spectra, but an extra line clue to some substance or other, Avhich has come into existence with the changed conditions in the nova. The three other lines mentioned above are probably the remnants of enhanced lines of iron which occur so prominently in the earlier spectra. March 28. Not very much different from March 27. The hydrogen lines, taken as a whole, show a complex structure. There is a broad bright portion divided in the middle by a rather indistinct dark reversal or absorption line. Then on each side of the bright line there is a rather narrow absorption line, the one on the violet side being far more conspicuous than the other. Then again on the outside of each of these absorption lines there are portions of radiation, but of considerably less intensity than that of the main radiation. The whole appearance of the hydrogen lines gives the impression of there being a broad intense bright line showing a black reversal, this bright line being nearly centrally superposed on a broad dark line, which again is centrally superposed on a broad bright line of less intensity than the central portion. The other lines remain about the same as on March 27. March 29. Thin photograph. No striking change from March 28. April 1. Very little different from March 28 and 29. There has appeared a rather broad, indistinct radiation in the position of the X 4471 '6 line of helium, but as there does not seem to be a corresponding radiation at X 4026 '3, the position of another helium line nearly always associated in stellar spectra with X 4471 ' 6, the helium origin cannot be unreservedly accepted. The structure of the complex hydrogen lines shows some slight modifications from March 28, but as this change seems more developed in the photographs of April 2 and 3, fuller discussion is reserved for the spectra of these dates. Apart from the hydrogen lines, the most outstanding line is still the broad one whose mean position is X 4639, and which seems to be of a dual nature, with the stronger and broader component on the more refrangible side. April 2. The hydrogen lines H/s and H 7 exhibit complex structure. The main portion is a double bright line, the less refrangible component being the brighter. The dark dividing line is very narrow in H/s, less so in H r . On either side of the strong bright H y line there is an absorption portion, that on the violet side being the broader. On the outside, again, of each of these bits PHENOMENA OF NEW STARS. 39 of absorption is a narrow bit of radiation, and yet again outside these pieces of radiation there are narrow bits of absorption. All the separate pieces mentioned are probably parts of a very complex structure in the 'hydrogen lines. The H> line shows somewhat different structure from that in H y The definite dual nature of the very bright radiation at H 7 is not seen in He. The least refrangible portion of the broad bright line is distinctly strongest, but the rest, up to the more refrangible edge, is of fairly uniform intensity, but split up by two rather indistinct narrow lines. The brightest, least refrangible portion of the bright line is fringed on the red side by a portion of radiation of less intensity, with a bit of absorption between the two. On the violet side of the broad bright line there is a piece of absorption, then another bit of radiation, then again a narrow piece of absorption, which in the reproduction is lost in what appears to be a broad uniform absorption line. Here again all the separate pieces of apparent radiation and absorption are probably part of a very complex structure of the H; line. The bands at XX 4176, 4233, 4471, 4-515, 4549, 4584 are present, but not very strong. The broad and diffuse line at X 4639 is still seen. The K line of calcium is very weak. April 3. Very similar to April 2. The structure of the hydrogen lines is still very complex ; Hr differing considerably from H r and H/s. The absorption lines to the left of H/s and H 7 , taken in their broad aspect, are divided by a bright line, not in the nature of a fine narrow reversal, the width of the bright dividing line being about the same as each of the two portions of absorption which the line divides. At H?, however, this broad absorption line is not equally divided by a radiation line, the more refrangible part of the absorption being a little broader than the right-hand portion and the bright dividing line combined. There seems to be a recrudescence of some of the lines which appeared on the red side of H 7 in some of the earlier photographs. Most of the enhanced iron lines are still seen in the nova spectrum. The spectrum in its main features is not unlike that of March 19, except that the enhanced iron lines are weaker in the later spectrum and there is considerable difference in the lines between H 7 and X 4515. The broad hazy band at X 4640 which has been in the spectrum since March 27 is here better defined, not so broad, and more centrally divided and very similar to that of March 19. The H line is a broad double, the less refrangible component being much the stronger. The K line is very weak and this is one of the most striking differences between the spectrum of this date and that of March 19. April 4. Not a very good photograph. The most notable change is that the prominent line at X 4640 has again assumed a broader and more diffuse appearance and, although double, is not symmetrically divided, the more refrangible portion being far the stronger. F 2 40 SOLAR PHYSICS COMMITTEE. April 5. Similar in its general features to the spectra of April 3 and 4. The hroad diffuse line near X 4640 now shows no indication of being a double line as it was on April 3 with a clear dividing line down the centre. The enhanced lines and X 4471 are still seen, but not prominently. April 6. The line near X 4040 is here a far more compact line than heretofore and it is very nearly as prominent as the hydrogen lines. It shows no suggestion of duality and it is easier to get a fairly accurate measure of its position than in previous photographs. Using the hydrogen lines as fiducial lines its wave-length has been calculated to be about X 4639. An extra radiation not very strong has now appeared about X 4G8, but whether this is a trace of the chief line X 468G in the series of hydrogen lines discovered by Prof. Pickering in Puppis is doubtful. The radiation at X 4471 and the enhanced iron radiations at XX 4176, 4233, 4515, 4549, 4584, are still shown weakly. An extra radiation at about X 4264 is here fairly well seen for the first time, though there was a trace of it in the spectra of April 4 and 5. April 7. The bright hydrogen lines now show the less refrangible portion much brighter and there is a dividing line of absorption, but not very distinct. The most striking change from the spectrum of April 6 is in the region near X 4640. The comparatively sharp and compact bright line of April 6 (X 4639) is replaced by a broad diffuse double as in the spectra of April 2 and 4. The more refrangible component of this double is much the stronger. The peculiar point is that the line of the 6th fits neither of the components of the double of the 7th. Neither does it seem to coincide with either the dividing line of the bright double nor with the middle of this double taken as a whole, although it approximates more closely to the latter position than to any one of the other three. The remnants of some of the enhanced iron lines (XX 4176, 4233, 4515, 4549, 4584) are still faintly seen. Also the radiations near XX 4264 and 4471. April 8. Not a very good spectrum. Prom what can be seen in it, however, there is no decided change from April 7. The lines in general are a little narrower than in the preceding photograph, but this may be accounted for by lack of exposure. April 9. Very similar to that of April 8, the line near X 4640 being a little narrower. April 10. Very similar to April 8 and 9, except that the line near X 4640 has further narrowed. April 13. A weak photograph. Only the hydrogen lines Up, H 7 , H,j and H seen, with a trace of the line near X 4640. PHENOMENA OF NEW STARS. 41 April 15. The spectrum is now almost a replica of that of April 6, in which the line X 4639 had assumed such a compact and comparatively well-defined appearance. The hright hydrogen lines show a similar structure to that of April 6. Also faint traces of XX 4233 (p. Fe), 4264, 4471, 4515 (p. Fe), 4584 (p. Fe), and 4680. April 16. The spectrum has again greatly changed in the vicinity of X4640 from that of the previous night, again showing the broad diffuse hright double line of April 7. This is the only very significant change, but there also appears to be a change in the relative intensities of some of the minor bands between X 4471 and X 4650. 4.^RELATION OF THE NOVA SPECTRA TO THE BRIGHT-LINE SPECTRA OF OTHER STARS. (a) WOLF-RAYET STARS. The spectra of a number of stars display bright lines mixed, in various ways, with dark lines which are frequently of known origin, and this has suggested to many observers the probability that some direct relations may be established between such stars and novae. The most numerous class of bright-line stars is that discovered by M.M. Wolf and Rayet, and studied at some length by Campbell, who has given a list of lines" ::: ' observed by him photographically and visually. In Table XI. we give in the first column all the lines recorded by Campbell in 31 Wblf-Rayet stars, and in the second column the number of stars in which each line was observed ; the signs + and -f + preceding certain XX in the first column indicate " bright " and " very bright," respectively. The data given in columns 3, 4, 7, 8 and 9 are also taken from Campbell's paper. The discussion of the origins embodied in the column for " Remarks " has been undertaken by Mr. Baxandall. Other celestial light sources are included in Table XL, so that no suggested relation may escape notice; the various headings to the several columns are sufficiently explanatory. * Astronomy and Astrophysics, Vol. XIII., p. 448. 42 SOLAR PHYSICS COMMITTEE. o .= 8 -g CO CM O o co S co 01- t:s -tl -^H ? 1 rT CO 2^ T fcC O C 2 8-- J s I- | 2^ "3 CO ^ ' X 1 9 2 5 J c a S *= o i . 31 . ^ o a S S o s> . CM * 1 CO CM CO o GO CM CO CM CO CM CM CO CM .- 5P vn c 1^ CD O CM O CO >O CM co CM CM CD CM t CO O CM O CO "= i = . = "* "o ^" T t '3 . S =' 8 1 1 = = 3=51 * | .2 o ^ o ^ = S ^ I -, >. .- 1 K .i "" 1 S to.2 C S ' s i a 5 s!j .! 1- ?|| j^^o r a C 4a Z~3 o "~ ~ ^ ^-** - !. *> K oo -~ S 2 s. r .2 .2 .5 o 1 i 1 113 x =i' 2 o -s g| S ,0 g .E ^ ^ 2 - 2 "5 - ^ - .2 -~ ^. l^ o ^ -4 1, ^ jfd 2 o~'S * ^ = ~f ~ w ~ 00 ~ = o^o "5>> = "- T x - be ^""^ S - ^^ "or *^ - ! *~ x ^ " "~ c ^ ~ ^o O ^ -^" * 3 C *^ C G ~ 1< o o 5: ^ ^ "TT i 2 "^ * "^^ ~ "^^ "T x = ?"l ? - = H ^ ^ ^ S o "^ " r r c *E "d^ 5 - 7 "* i "fr -^ ^ .- 7 i "* 8 I c ^c ^ ?j" c ~ ~ * -^ w ' ^ i 5 TT X ~ ^ ^^ ^^ ^ ^ | _ MM ^ - ^ Sj . r v ^ < ^5 i ~ ^ = - "* S d "* -S ""xo .E! . _. = 5 - C |~ Z >~. 5 , ~ = 5 2 ., , ~ ^ o"""*"^ S^" 2^ : 1^11 1=;H |. i *i : Z 52 ^ - Z.-3 ^ "= ~ ~ = S i = * "i"l~ s - a 06 * J "i -5 = i ~ a! = 2 2 -f = J S 11 - " c 2 =5^^! ji^ ^-7 ^ ! i ^ I - I =*^ ^" ; ~ -. i ' '= S * n . 2 "r J- "" J* c i ~ ?- s _2 iT* ' "s 's E /-N "" r ' ^ s~ ^~ x -~ hU >H M ^ , a o -* a -0 - ^^=~ X 1 1 | 1 1 | 1 IS 1 1 1 71 1 1 r- x TJI | 1 1 -r -r T-- M ^ IM 1 1 *? 1 55 1 1 so II \ ^ -! -r * 52 7-1 -^ O 1- O 55 M X 1 4 i 1 i eo 5 t ^i* "T" ~T ~^ ""f -^ i -* 71 - i o o - * I ** a> <- n ~ co 7^ CO O t- X ~. -T -! -^ 11 C X O CO - o - 3 1 -* t~ HO ~. \ # II *** II- II = 1 S s 2 -T' ^ X wl O ~ 7-1 | o | ~ * e, 7-) x r. ~ ro i O rt O ^~ -r- >C S 1 X a 1^1 -T + + 44 SOLAR PHYSICS COMMITTEE. >-. X 0) h-5 'C a cu a> ^H O CD o g ^ Si _9 J| >f 'u 03 - fc. f 2 "5 _ - o 2 a 01 ^ S 10 -=. c S o S -g - a> o -^ fc, CM CO S tT 5 S? S 85 g S > "? v_^_ r^ P^H T" CP w oo 'o -2 111 1 ~~z. ^ 5*"" Yji o o so ^ ri J .22 ^ -S o 'E a j3 "T i2 3 S 02 J S .s '5 i O 03 2^ ^ j^i S - O r >xj- O ^ ~ w c? "s .3 s o , ? O ^f^ "g - fee S 2 "* 5" a's*i< a ^ I "1 ^ Tr ' -" o t? a? '3 a .- S C Sf > t~ ^ S -w 6 = 5 g o >^ g j^ "2s"" 1 *' ,5 co a ' *2 J3 H *T ' ' r ^ ^ ^ ^ p CO -2 ^ > M "^ o ^3 >> * ^4 ^ C3 P ^- '^ ^> *"Q "^ ^"^^X 3D g c -7 7 5. s ^ ^ p^"&S ^ il^lj .2 "S s Mi 1 1 s ii 111 1 1 I |J 'S "t-tc o OO ^ O o o & 2 '- CO e5 "3 S C CO "p S 1-1 ^" ^i PH M &H M " R - I 11 C -f CO CO CO ' a v - / i-^ w. hJ O ^3 -3 ^ 'O 4 O y_,J "** , EH Sffi 3 c ^ /-N t^ Tfl O O X 00 CO X i CO 'f CD 00 CM O Q) G Q O *O 'O 1 O O ^O CO **& "* "* "* "* -* -* Cj ^Jj^j o'C .2 r- co I O I o [ 85 1 o 5 CD s. 2 o & G '~ CO r CO O m S a, O > 10 ^3 co CO 2 o -2 u G ' OO 05- 10 co >o c *"' ^~ co o ! , O O * O CO >-O O CO CJ1 CM CO "O CO CO CO -3 < + + + + PHENOMENA OF NEW STARS. 45 t> c3 = "so o ** a - O - S ^= S ~ .2 < - = ** O QJ .b 2 O j.= ; *3 x ? -5* -S ills IIJ1 l * S 2 no ~ = c~2 = f u ~ s~ - c 8| I S > t ( O * a = ? O O 00 0, 2 '- 03 e ~ -S-5 .23 s s 2 .3 I *4-t O s a., 'vf O > >-, a w = = 3 ? c =: s >> o 3 .= p 2 .- '-5 ? g 2 S a 1 CM 1C o -t> I I 1 i i CN O CO X o O OJ CM X * CO 00 * IN Si * CO J~ CM I I I M to g I I to be 31 O S 'be g 03 PH o S s O> be o co rf-t >O ,= o o * o 1 i I ,0 c rO Q* S X C5 30 1C 00 CO >O O >^ co i- o i?. CO X CO co 00 at 00 co CO CO >o X CO r~ XI CO co CO rM CO CO CO CO OJ i CO 00 >O c 00 CO >O CO PHENOMENA OF NEW STARS. 47 From this table and the accompanying remarks it is obvious that the spectra of Wolf-Rayet stars and novae in the bright-line stage have very little in common beyond the hydrogen lines. But the Wolf-Rayet stars are essentially gaseous masses from which the metals, in the form producing the enhanced lines of the nova bright-line stage, have been eliminated, so far as radiation is concerned, and therefore we must naturally look for similarity to the Wolf- Rayet spectra in the spectra of novae in the X 4640 and nebular stages, where we find gaseous bright lines only and none due to metals. Under such conditions temperature alone probably plays a less, and electrical effects a greater, part in the production and variation of radiation. In Table XII. we give a comparison of Campbell's Wolf-Rayet spectra with the spectra of novae in these later stages. Column 1 contains the Wolf-Rayet lines published by Campbell,* with a frequency number showing in how many stars of the 31 studied by him each line occurs. Column 2 contains Campbell's results! for Nova Aurigae when in the nebular stage, and 3 the spectrum of Nova Persei as photographed and reduced at South Kensington. + The wave-lengths given in column 4 for Nova 'Geminorum (1) comprise' two sets, one published by Perrme and the other by Messrs. Reese and Curtis, I ; the latter are given '.to four figures. In column 5 are given the Avave-lengths of the spectrum of Nova Lacertae published by Dr. Wright^ and derived from photographs taken, at the Lick Observatory, when the nebular line X 5007 was the strongest line in the spectrum. All the preceding columns deal with the nebular stage only, the X 4640 stage not having been treated separately in previous novae. But in the sixth and seventh columns we give two sets of wave-lengths, the first showing those of the few bright bands in the Madrid spectrum of Nova Geminorum (2) when in the X 4640 stage (see p. 32), measured at South Kensington, and the second those of the same nova when in the nebular stage as recorded by Stratton ::::S ; in 'the latter the numbers in brackets show the order of intensity of the bands. * Astronomy and Astrophysics, Vol. XIII, p. 468. t Ibid, Vol. XII., p. 726. + Proc. Roy. Soe. Vol. 69. p. 135, 1901. Astrophysical Journal, Vol. XVIII., p. 297. I Ibid., pp. 299-306. ^ Lick Observatory Bulletin No. 194, p. 100. ** Monthly Notices, Vol. LXXIII., p. 72. G 2 48 SOLAR PHYSICS COMMITTEE. TABLE XII. COMPARISON OF THE SPECTRA OF WOLF-RAYET STARS AND OF NOV^E IN THE NEBULAR STAGE. 2 Wolf-Rayet Stars (Campbell). Fre- quency of Occur- rence. Max. = 31. Nova Aurigae (Campbell), August 1892. Nova Persei (South Kensington), August 26, 1901. Nova Geminorum (Perrine) and (Reese and Curtis), April 1903. Nova Lacertaj (Wright), March 30, 1911. Nova Geminorum (2). Origin. (South Kensington) 4640 Stage, April 15, 1912. (Stratton), Nebular Stage. Dec. 6, 1912. Int. Int. AA AA Max. AA Max. AA Int. AA AA AA = 10. = 10. 384 H, 3868 10 H 389 391) tr. 3970 8 397 _, 3970-2 H. 1 3982 4023 1 He 4063 3 _ 4090 4102 4 4098 0-2 _ 410 4101-8 Hi (7) Ha 420 4 4228 3 423 0-1 ' 4260 2 426 0-1 ... .. _ 4273 4 , 4318 5 ; 4334 . 1 . , 4341 9 4336 0-1 4341 2 434 v.s. 4340-6 H y (5) H, 4342 4369 4 4358 0-8 4364 7 1 1 436(1) 4378 __ _ 1 4389 2 438 0-1 A st. 4416 1 4443 10 4457 1 4466 4 1 4473 2 4466 0-1 4471 1 446 4470-2 He 4480 5 p. Mg PHENOMENA OF NEW STARS. 49 COMPARISON OF THE SPECTRA OF WOLF-RAYET STARS AND OF NOV.E IN THE NEBULAR STAGE continued. 1 2 3 4567 Wolf-Rayet Stare (Campbell). Fre- quency of Occur- rence. Max. = 31. Nova Aurigae (Campbell), August 1892. XT . Nova Gemi Nova Persei (South (Perrine) . (Reese a Kensington), August 26, 1901. April 19 aorum Nova Nova Geminorum (2). and Lacertae , (South (Stratton), nd (Wright). Kensington)! Nebular Origin. >, March SO. 4640 Stage, Stage, m 1911 April 15, Dec. 6, 1912. 1912. Int. Int. AA AA Max. AA Max. x*. Int. AA AA AA = 10. = 10. 1 4493 2 _ 4504 ) 8 451 0-1 4517) 4536 4541 19 p. II 4555 1 4596 o 460 0-1 4615 6 4626 5 462 4636 10 4630 0-7 4636 3 4630 4639-2 464(3) Unk. 4652 16 4656 4688 31 4681 0-4 4684 2 4680-2 468 (8) p. H 471 0-1 4720 6 471(9)1 4787 14 4802 <1 -* 4862 22 4857 1 4862 3 486 *.n. 4861 4861-49 H,| (4) H s 4861 4940 15 . 4953 3 4959 2 4959 496 (6) \ 2nd neb. 5002 10 5007 4 5007 5006 501 (2) 1st neb. 5020 14 5131 13 i 5250 6 _ 1 5284 3 : 1 5412 24 - p. H 5472 13 5557 5593 15 5671 ~~ ~ ~ SOLAR PHYSICS COMMITTEE. COMPARISON OF THE SPECTRA OF WOLF-RAYET STARS AND OF IN THE NEBULAR STAGE continued. 1 2 3 4567 Wolf-Rayet Stars (Campbell). Fre- quency of Occur- rence. Max. = 31. Nova Auriga; (Campbell), August 1892. Nova Persei (South Kensington), August 26 1901. Nova Geminorum (Perrine) and (Reese and Curtis). April 1U03. Nova Lacertae (Wright), March 30, 1911. Nova Geminorum (2). Origin. (South Kensington) 4640 Stage, April 15, 1912. (Stratton), Nebular Stage. Dec. 6, 1912. AA AA Int. Max. = 10. AA Int. Max. = 10. AA Int. AA AA AA 5693 18 5739 5750 5752 5813 21 I 5848 1 . 5877 9 . D 3 6564 5 6563 H. As most of these wave-lengths in Tahle XII. are connected with the nebular stage we deal with that first, leaving the precedent X 4640 stage till later. At first sight the table conveys the impression that although there are more points in common between the Wolf-Rayet stars and novae in the " nebula " stage, than between the former and the enhanced-line stage of novae, the agreement is not conclusive. The hydrogen lines are common, but apart from them there is no known element certainly represented in both types of spectra. The helium line X4471 appears in most of the novae, but it is not accompanied by the other helium lines, such as A. 4026, the next line of the same series, and, so far as has been recorded by Campbell, only appears as a bright line in two of the thirty-one Wolf-Rayet spectra, although it was photographed as a dark line in three others. On the other hand the visual observations revealed the presence of D 3 in nine stars of the Wolf-Rayet type, and in all there is evidence, more or less complete, for the presence of helium in 12 of the 31 stars observed by Campbell ; but only in three stars is helium represented by more than one line, and in one of these the second line (X 4471) is dark while the other (D 3 ) is bright. We cannot, therefore, definitely claim helium as a common connecting element between novae and Wolf-Rayet stars. As regards proto-hydrogen, Rydberg's principal series, as represented by the X 4688 line, appears in every Wolf-Rayet spectrum examined by Campbell, and is possibly represented in the "nebula" spectrum of Nova? Aurigae, Persei (2), and Geminorum (2) ; the other lines of this series, recently photographed terrestrially by Powler,* are too far in the ultra-violet to be shown in photographs of stellar spectra. The absence of the other p. H lines, the Puppis series, is not conclusive evidence against the p. H origin in the novae, for the two lines X 4542 and X 4201 * Monthly Notices, Vol. LXXIII., No. 2, December 1912. PHENOMENA OF NEW STARS. 51 are only represented nine and four times, respectively, in the Wolf-Rayet spectra, and the X 4688 line, as Rydberg has pointed out,* is much the strongest line of all the hydrogen series and might, therefore, be expected to reveal itself when the other p. H were too weak to show; the ordinary hydrogen lines are considerably weakened when the novae spectra reach the nebular stage. From this evidence we conclude that the connection between the spectra of novae in the nebular stage and those of Wolf-Rayet stars in general is almost entirely dependent upon the identity of the ordinary and the Rydberg principal series of hydrogen. But if we consider specific Wolf-Rayet spectra in connection with the X 4640 stage of novae, we find a stronger resemblance between the two classes of stars. Before doing this it should be remarked that Campbell's list of lines given in Table XII. is a summation of a large number (31) of the star spectra in which the individuals differ considerably inter se, and that many stars of the Wolf-Rayet type" could not be observed at Mount Hamilton. The specific Wolf-Rayet spectra that are similar to the spectrum of the novae in the X 4640 stage are to lie found among those of Miss Cannon's O class (Argonian) stars. Thus in the spectrum of the typical star ( Carime A.G.C. 15305) of the Oa class, a star too far south for Campbell to deal with, Miss Cannon finds t that " a broad bright band, whose centre is at the wave-length 4633 is the " most conspicuous feature of this spectrum. On the side of shorter wave-length " the edge of this band is well defined and resembles a dark line ; on the side " of greater wave-length, the brightness fades off into a fainter band of nearly equal " width, which may coincide with band 4688 seen in classes Ob, Oc, Od, and " Oe." The general spectrum consists of bright bands on a faint continuous background, no dark bands being present. In addition to the two already named the only other bands in the spectrum are He, Hy, and possibly the X 4471 helium line ; the Puppis series of hydrogen lines may l)e there, but are not seen with certainty on any photograph. A comparison between this spectrum described by Miss Cannon and the spectrum of Nova Geminorum (2) in the X 4640 stage (given in Table X.) is made in Table XIII. on the following page. In making this comparison it must be recognised that very few lines are involved and any conclusions drawn from it may therefore appear somewhat doubtful. But on the other hand the few lines are an outstanding characteristic of both spectra, remarked by Miss Cannon in the case of the Oa type-star and obvious in the case of the nova [Geminorum (2) ] on reference to Table X. It is also important to note that in the Oa spectrum the bright bands appear on " a faint continuous background," and reference to the Madrid spectra (Plate II.) shows that . when (April 15) the 4640 stage is arrived at, the con- tinuous spectrum has almost disappeared Visual observations on April 11 and 19 (see p. 24) also revealed the faintness of the continuous spectrum. * Astrophysical Journal, Vol. VI., p. 237. t "Spectra of Bright Southern Stars," H.C.O. Anuals, Vol. XXVIII., Part II., p. 146. 52 SOLAR PHYSICS COMMITTEE. TABLE XIII. NOVA GEMINORUM (2). X 4640 STAGE, APRIL 15, 1912. SOUTH KENSINGTON. AX Orig. Remarks. 3970-25 H Middle of bright H, (3 weak lines.) 4101-85 H Middle of bright H 4 (3 weak lines.) 4340-66 H Middle of bright H y 4470-2 He Middle of broad weak bright band. 4639-2 Middle of strong isolated bright line. 4680-2 ?p.H Middle of weak broad line, diffuse to measure, possibly 4688 of Rydberg's H series. " 4861-49 H Middle of bright H,, TYPICAL Oa STAR (- CARING, A.G.C. 15305). HARVARD. xx 4059 4101-8 4340-7 4471-8 4633 4688 Orig. H H He Remarks. Broad bright band, most con- spicuous feature in spectrum ; fades off into fainter band which 'may coincide with 4688. Also the Puppis hydrogen lines are suspected as bright, but are not clearly seen on any photo- graph. Thus we find a general similarity between the spectra of the X 4640 stage of novae and the Oa (Wolf-Rayet) spectra, a similarity which extends to practically all the features exhibited by these spectra. Prof. Fowler has shown that to get the p. hydrogen lines showing clearly in laboratory spectra, electrical discharges of great intensity are necessary* in the vacuum tube. In the novae spectra the special conditions are clearly not fully attained, but the presence of the X 4688 radiation in both novae and the Oa spectra, and the possible presence of the bright Puppis series in the latter, indicates the probability that in the atmosphere of these stars the laboratory conditions are approached. With regard to the apparent variation of the wave-length for the " X 4640 " radiation as given by different observers in different stars, it must be borne in mind that the line is always broad and sometimes diffuse and therefore difficult to measure. Further, many of the measures have been made visually or on photographs of very small dispersion and they must therefore be held to be liable to large probable errors. In the type-stars, Miss Cannon gives the value as 4633, in y Velorum (Oa Pec.) the mean wave-length is given as 4635 '7, while Copeland's visual measures at Vincocayat gave 464 6 and our recent measures of the band in Nova Geminorum (2) gave 4639 '2. It is, however, -probable that the same radiation was measured in each case. ' Monthly Notices, R.A.S., Vol. LXXIIL, No. 2, December 1912. t Copernicus, Vol. III., p. 205. PHENOMENA OF NEW STARS. 53 In relation to novae, which show simultaneously bright and dark lines in their spectra, it is interesting to note the features which differentiate the spectra of the several groups of Pickering's O class. This class has, in the Kensington classification, been treated as a whole and is designated " Argonian," but in the Harvard classification it is subdivided into five groups, Oa to Oe inclusive ; a similar subdivision of the " Alnitamian"" class has recently been made at South Kensington.* In the Oa class, as we have seeen, the radiation X 4033 is the strongest and occurs in a spectrum of bright lines only, the continuous spectrum being weak ; X 4688 and, possibly, the Puppis hydrogen lines are also present. In the Ob class the X 4633 hand is absent, X 4688 is very intense and the Puppis hydrogen lines are present ; again the spectrum is purely a bright-line spectrum. In Oc X 4688 is still the most conspicuous feature, but X 4633 is present with the additional hydrogen lines, and no dark lines are shown. Class Od ( Puppis) shows all dark lines except X 4688 and X 4633, but. the spectrum contains only four lines in addition to the two series of hydrogen. On the other hand, the Oe class contains many dark lines in addition to XX 4633 and 4688, which are bright. By an intermediate stage (the OesB class) the Harvard classification passes to the purely dark-line stage represented by e Orionis (B, or Alnitamian, type) in which XX 4633 and 4688 are not represented and the Puppis lines are fainter. Thus, we see that in the Argonian stars, as in novae, the X 4633 bright line comes out as a prominent feature approximately at that stage where bright lines are associated, in some form or another, with dark lines ; this, also, suggests a similarity between novae in the " X 4640 " stage and Wolf-Eayet stars. There is another suggested relation between Wolf-Rayet stars and novae which should not be overlooked. Prof. Hartmann found, on examining a photo- graph of the spectrum of Nova Persei, taken on October 15 and 18, 1907, that " das notaspektren fast voUkommen identisch mit clem spektrum des Wolf- Eayet-Sterm BD 35 4001."t This, coming after the nebular stage, at first suggested the possibility of a reversion in the spectral changes of the nova. The star BD 35 4001 is of a fairly typical "Wolf-Rayet character,;}: its spectrum, according to Campbell, including the radiations XX 4862, 4650, 4627 (H ), 4598, 4541 ( + ), 4508, 4481, 4465, 4442, 4369, 4341 ( + ), 420 ( + ), 4102 ( + ), and 406. In both the nova and the Wolf-Rayet spectra the line 4687 was the strongest, but the small scale of the photographs used by Prof. Hartmann made it impossible to go into any detail. In answer to an inquiry Prof. Hartmann very kindly informed me that the two spectra displayed a general similarity in the distribution of radiation, the * On the Spectra of the Rigelian, Crucian and Alnitamian Stars. Solar Physics Committee Publications, 1914. t Astronomische Nachrichten, No. 4232, coL 115, Vol. 177. J Astronomy and Astrophysics, Vol. XIII., p. 462. a 19881 H 54 SOLAR PHYSICS COMMITTEE. maxima occurring as stated in his original note, hut further than that he was unahle to go owing to the faintness and small scale of his spectra. Thus, while it is ohvious that hetween the winter of 1901 and that of 1907 some marked change occurred in the spectrum of Nova Persei, we have not sufficient data to enable us to fix definitely a stage in the spectrum, or a physical condition, which would hri'ng the nova at that epoch into line with the other phenomena attendant upon novae. 4 (i). OTHER BRIGHT-LINE STARS AND In addition to the Wolf-Rayet stars, there are a numher of stars giving hright radiations in their spectra. In some cases absorption lines accompany the radiation lines, in others the dark lines are sporadic, while in other examples, e.g., /3 Lyra;, the relative positions of the dark and hright companions change about from time to time. These stars are not very numerous among the brighter objects, and their spectral phenomena beyond the existence of bright lines vary so greatly from star to star that it is almost impossible to classify the objects in more detail. But although the phenomena are so varied and the objects relatively few, the problems presented are probably among the most important in astrophysics. In the following table (XIV.), we bring together the bright lines recorded in the spectra of several such objects, and compare them with the spectra of Nova Geminorum (2) in the " bright-line " and the " A. 4640 " stages, and the spectrum of Nova Persei (2) in the nebular stage. It is necessary to take two novae into consideration, because in the case of Nova Persei the X. 4040 stage was not considered specifically, while in the case of Nova Geminorum (2) we have not yet a complete analysis of the nebular stage. TABLE XIV. BRIGHT LINES IN THE SPECTRA OF NOV^E IN DIFFERENT STAGES, AND OF BRIGHT-LINE STARS. N. Geminorum (2), /3 Lyne Madrid Spectra, N. Persei (2) South Kensington reduction. (South P. Cygni (i Centauri (Pickering Kensington) (Frost). (South and Remarks. " Bright-line " "A. 4640" Nebular Kensington). South stage, stage, stage, Kensington.) Mar. 19, 1912. Apr. 15, 1912. Aug. 26, 1901. AA AA AA AA AA AA 342 von Gothard's nebular line. . . 3819-2 3856-2 . 3868 3889 3887 H ? + He. 3934-6 3933 p. Ca. 3964 He. 3969-2 3970-25 3970 3970 . H + II (Ca). 3995 N. 4006-9 v, PHENOMENA OF NEW STARS. 55 BRIGHT LINES IN THE SPECTRA OF NOV^E IN DIFFERENT STAGES, AND OF BRIGHT-LINE STARS continued. X. Geminorum (2), ft Lyra Madrid Spectra, N. Persei (2) South Kensington reduction. (South P. Oygni ft Centauri (Pickering Kensington), , (Frost) . (South and Remarks. Bright line " ' A. 4640 " Nebular Kensington) South stage, stage, stage, Kensington). Mar. 19, 1!12. Apr. 1, 1912. Aug. 26, 1901. AA AA AA AA AA. AA 4019-3 4026 4026-4 He. 4061-0 < 4081-5 Apparent reversal iu dark H 8 . 4101-85 4101-85 4102 4102 4101-8 4101-8 Hi 4121 4120-5 He. 4143 4144-0 He. 4154-4 4174-5 4176-4 p. Fe. 4232-5 4232-9 p. Fe. 4252-8 4267-2 ?p.C. 4301-1 4299-4 p. Fe. 4317-3 . . 4340-66 4340-66 4341 4340 4340-6 4340-6 HT. 4349 0. 4351 0. 4357-9 Narrow radiation bor- dering H y . Probably the less refrangible por- tion of a broad line, part of which is hidden by bright H r this broad line being probably identical with the a Cvgni p.Fe line X 4351 -9. 4364 4367 0. 4376-4 . 4385 p. Fe. 4388 4387-8 He. 4392-3 4408-7 4419 4423-8 4434-2 4450-4 4470-0 4470-2 4471 4472 4471-8 He. 4508-9 p. Fe. 4516-4 4515-1 p. Fe. 4523-3 - 4549 4549-9 p. Fe. 4556-3 4584 ' 1 4584-6 p. Fe. 4491 0. 4601 N. 4607 N. 4622 N. 4631 N. 4633-1 p. Fe. 4639-2 4636 . 4643 - N. 4660-2 4661 O. 4676 O. 4680-2 H 2 56 SOLAR PHYSICS COMMITTEE. BRIGHT LINES IN THE SPECTRA OF NOV.E IN DIFFERENT STAGES, AND OF BRIGHT-LINE STARS continued. N. Geminorum (2), N. Persei (2) Lyras Madrid Spectra, (South t South Kensington reduction. Kensington.) P. Cygni H Centauri (Pickering Kebular stage, (Frost). (South Kensington) and South Kensington.) Remarks. " Bright line " stage. " A 4640 " stage, Mar. 19, 1912 Apr. 15, 1912. Aug. 26, 1901. AA AA AA AA AA AA _ 4684 4698 ' 1 4713 4712; 8 He. 4720 4727-8 4808 4861-49 4861-49 4862 4861 4861 4861 H d. 4'923 4923 7 He. 4959 5007 5016 5023 He. 5055 5127 . 5155 5240 5876 He (D,). 6563 H . ~ 6678 He. This comparison does not show any striking likeness between any of these typical bright-line stars and novae. The hydrogen lines are common to all, and so is the X 4171 helium line, excepting yu. Centauri, but many other helium lines which appear in P Cygni and ft Lyrae are absent from all the novae spectra here dealt with. The spectrum of /a Centauri does resemble that of the novae in the bright-line stage, in that many of the lines are due to proto-iron, but they are not always the same lines. No analogue of the nova X 461-0 stage presents itself, and the entire absence of the typical nebular lines from the bright-line spectra considered, absolutely debars any analogy being drawn between these spectra and the nebular-stage spectrum of the novae. Thus, we arrive at the conclusion that the physical conditions obtaining in novae, in any of the stages pictured in their spectra, are very dissimilar to those obtaining in the typical bright-line stars considered in the above table. 5. THE STRUCTURE OP THE BANDS IN NOV./E SPECTRA. One of the most striking features of the spectra of novae, especially when in the " bright-line " stage, is the multiple character of the broad bright bands, particularly those of hydrogen, such as might be expected if we were dealing with a complicated system in which the different units were moving in the line of sight with enormously different velocities. PHONOMENA OF NEW STARS. 57 In most novae spectra this phenomenon has been remarked, and in the case of Nova Aurigse the strange and apparently varying multiplicity was so puzzling that Schumann pointed out" :;: " that if a multiple system were invoked to explain it, the system would have to be at least sextuple. Belopolsky's results, quoted by Vogel,| required an even greater complexity to explain them. In the case of Nova Persei (2), many intensity curves of various lines, principally of hydrogen, were published. The South Kensington intensity curves appeared in several papers communicated to the Royal Society in 1901,+ and showed numerous changes in the dispositions and intensities of the maxima seen in the hydrogen lines. If these varying displacements be taken as a Doppler- Fizeau 1 effect, the relative velocities shown by the extreme bright maxima of H/3 amounted to as much as 1,000 miles per second, while the relative velocity of the two central maxima was about 200 miles per second. It is to be noted that these broad, bright bands generally do not fade in intensity from one edge to another, but usually have sharply denned edges on both sides, indicating that the motions, if motion can be held accountable for the shifts, are those of well- defined streams of matter, each stream moving with its own peculiar velocity. In the case of H 7 the early photographs showed three maxima, the two outside ones being broad, the middle one narrow, although almost of equal intensity ; the curve resembles the sharp double reversal such as is found in the solar calcium lines H and K in the neighbourhood of a solar disturbance. Between February 25 and March 10, 1901, the maximum intensity in this band changed from the more to the less refrangible side of the bright band. In the case of Nova Geminorum No. 2 this complicated structure of the bright bands, in the bright-line stage, was again a pronounced feature, and the oscillations of the maxima within the bands was remarked by many observers. The changes exhibited on the Madrid spectrograms are discussed in Section '3 (J), and it should be noted that the phenomena were shared by the bright metallic bands, similar structure appearing on the same date and all varying together from one date to another, thus showing the complex chemical nature of the matter which produced the radiations. As the spectra of novae change from the " bright-line " stage to the later stages, the complicated structure tends to disappear with the subsidence of the disturbance until, in the " nebula " stage, the bands are much more homogeneous. In addition to the complex bright lines in the " bright-line " stage, there are also the dark companions to the hydrogen lines to be considered. These do not show the same structure as do the bright lines, but do suffer changes in apparent positions and appearance. At times they are not separated from the bright bands, and the suggestion has been made that their less refrangible * Astronomy anil Astro-Physics, Vol. XII., p. 159. | Ibid., Vol. XIII., p. 56. + Koy. Soc. Proc., Vol. 68, pp. 143, 144, 233, 234 and 404. 58 SOLAR PHYSICS COMMITTEE. edges might be masked by the emission, thus making it impossible to determine the wave-lengths of the centres of the dark bands ; but at other epochs there has been a distinct space between the less refrangible edge of the dark, and the more refrangible edge of the bright, bands. Examples of this are shown on Plate II., where the division between the bright and dark components may be seen on the photographs of March 25-April 5. 6. MAGNITUDE OBSERVATIONS OF NOV^l. i The phenomena attending novae in their early stages are so obviously cataclysmic in character that one would a priori, expect strongly-marked oscilla- tions in the general light-emission. Such oscillations have been observed in the magnitudes of those novae which have been discovered immediately after their first appearance during the early stages of their spectral changes. The light-curve of Nova Aurigse was a very remarkable one. Less bright than the eighth magnitude on December 8, 1891, it was of about magnitude 5'0 when discovered by Dr. Anderson on February 1, 1892, and then, after reaching a maximum, it declined gradually, with many intervening brightenings on a small scale,* until on April 26, 1892, it had sunk so low as the sixteenth magnitude,! according to Prof. Burnham's observations with the 36-inch refractor at the Lick Observatory. But observations made at the same observatory on August 17, 1892, showed that the nova had brightened considerably, and was then of magnitude 10 '5, while two days later it had become 9 '8, where it remained steady for a long period. Coming to Nova Persei (2), observations, made at South Kensington from February 25 to May 5, 1901, were plotted on a curve,J and showed that the magnitude was an oscillating quantity. From this and from the much more extensive data collated by Prof. Pickering it is seen that while the increase of light to the primary maximum took the form of an. enormous uprush, unbroken by any fluctuation, the decrease was relatively much more gradual and was marked by very frequent oscillations. In a communication to the Royal Society in June 1901 \, I pointed out that the spectrum of the nova appeared to vary in sympathy with the changes in magnitude, and this was confirmed by several other observers. Prof. Pickering showed that during the period March 17-April 28, 1901, the spectrum varied from " normal " to " nebula," as the magnitude changed Monthly Notices, Vol. LIL, pp. 357-368, March 1892. | U>id., Vol. LIIL, p. 59. f Proc. Koy. Soc., Vol. 68, p. 399.1 Annals of Harvard College Observatory, XLVIII, Part. II., Plate I. || Proc. Eoy. Soc., Vol. 68, p. 403. PHENOMENA OF NEW STARS. 59 from maximum to minima respectively, until, on the latter date, it became constantly " nebula " ; his " normal " spectrum is that which we have designated the bright- line spectrum.* Nova Lacertae showed similar light changes, but the oscillations recorded were neither so abrupt nor so frequent.t In the case of Nova Geminorum (2) the primary fall from the initial maximum was of the same sudden nature as that observed in the case of Nova Persei, and there were numerous oscillations, roughly synchronous with changes in the spectrum, during the two months following the outburst. As will be seen from the light-curves plotted from the observations published by Dr. Freundlich,+ the nova of 1911-12 did not exhibit the same regularity of period as shown by Nova Persei, but there is a general similarity. The curves shown on Plate III. have been plotted from the observations published by Prof. Pickering for Nova Persei (2), by Dr. Freundlich for Nova Geminorum (2), and Prof. Nijland for Nova Lacertae. Thus we see that the light-curves of probably all novae observed during their early stages are of the same general character, in which an enormous outburst, frequently multiplying the emission of light many hundredfold, is followed by a very abrupt subsidence, which after a day or two becomes far more gradual and subject to slight periodical recoveries. The different times taken, in different novae, for these changes to take place suggest that greatly differing masses, or velocities, are involved, while the rapid incandescence and subsidence in all cases point to the action of small individual masses and not to the collision of only two large bodies. 7. LOCATION OF NOV^E IN SPACE. When considering the location of novae in space we have to deal with two co-ordinates, the first giving us the distance from the solar system, the second the part of the apparent celestial sphere in which the nova appears. For the first we have but a very meagre amount of data to work upon, for only in two cases has the object teen near enough for its distance to be measurable with the present accuracy obtainable in parallax determinations. Nova Persei, in 1901, apparently offered the best opportunity, and the four values published, by different observers, vary from + 0'16" 0'06" (Hartwig) to - 0'012" + 0'035"; it is probable that the parallax did not exceed O'lOO". * Auuals Harvard College Observatory, LVI., No. III., p. 62. t Astronomische Nachrichten, No. 4562. (Prof. Nijland.) J Astronomische Nachriehten, No. 4624. 60 SOLAR PHYSICS COMMITTEE. An attempt to determine the parallax of Nova Lacertae, 1910, made by Dr. Slocum, with the Yerkes 40-inch refractor, gave the value + 0'013" +0'014<", and reducing this for the assumed parallax of the comparison stars a probable value, -f-0'018", was obtained for the absolute parallax of the nova.'* Franz found a large negative value - 0'32" + 0'12" for the nova| which appeared in the Andromeda nebula in 1885. These results and those obtained for other novae, only serve to point to the fact that, as a class, new stars are exceedingly remote from the solar system. The second co-ordinate is much simpler to determine, depending only on direct observation of the nova's apparent position, and the result is very striking. All novae, with the one exception of Nova Coronae in 1866, have appeared in the Milky Way, its branches, or the Magellanic Clouds. This preference for the Galaxy is well shown for the northern hemisphere on Plate IV., where the positions of all northern novae have been plotted on a photographic copy of Boeddicker's general map of the Milky Way. Such novae as have appeared since 1891 are marked by the number given in Table I ; those that appeared previously are marked by the name of the constellation and the date of their discovery. ADDENDUM. Subsequent to the completion of that part of the memoir dealing with the bright-line spectrum, and when the memoir was nearly ready for press, we received a paper by Prof. Adams and Mr. Kohlschutter discussing their observa- tions of the spectrum of Nova Geminorum (2) at Mount Wilson. It was then too late to deal with this paper in the body of the memoir, but it is thought desirable to publish the following table, even without fuller discussion, comparing the lines and origins found by the Mount Wilson observers. In the first column are given the wave-lengths. Messrs. Adams and Kohlschutter give a great number of wave-lengths for the separate parts of the bands, but in this table we have taken only such as are described as belonging to the centres of bright bands, i.e., the typical bright bands of the " bright-line " stage. The second column contains the origins ascribed to these radiations. In the third and fourth columns we give the wave-lengths and origins found at South Kensington from the dispussion of the Madrid spectro- grams of the same nova. Then follow the similar data we published from this observatory in 1902,J in which it was shown that these bright radiations of *Astrophysical Journal, Vol. XXXV., p. 137. t Astronomische Nachrichten, No. 2816. } Proe. Roy. Soc., Vol. 69, p. 356. Astrophysical Journal, Vol. XXXVI., p. 293. PHENOMENA OF NEW STARS. 61 Nova Persei, other than those due to hydrogen, are chiefly due to proto-metallic substances, that is to say they are enhanced lines. In the seventh column are given the wave-lengths of the corresponding enhanced lines and origins, and, speaking generally, they are the strongest enhanced lines, while finally we give the corresponding lines with their intensities in the spectrum of a Cygni. The wave-lengths given by Messrs. Adams and Kohlschutter are the observed wave-lengths, and are subject to corrections for displacement which for the hydrogen lines H f . B>, H 7 and H/s are -1'8, -1'3, -1'8 and -1'8 tenth-metres respectively. ADAMS'S AND KOHLSCHUTTER'S SPECTRUM OF NOVA GEMINORUM (2) COMPARED WITH THE SOUTH KENSINGTON RESULTS. A. & K., Nova Geminorum (2)- S. Kensington. N. Q-eminorum (2). 3. Kensington, N. Persei (2). S. Kensington, Enhanced Lines. S. Kensington, a Cygni. Remarks. XX Observed. Origin XA (Origin XX Origin XX Origin XX i Int. Mai. = 10. Origin 3891-3 Hf 3S89 Hf 3889-1 10 H 3955-2 3952-07 p. V 3952 1 1-2 p.V 3970-7 H ( : 3969-2 I j H ' . 3976-3 f * 3969 H.i Ca 3968 6 10 H 4035-0 -- 4035-80 p. V 4035 8 2 p.V 4067-9 Neb. 4067 p.Ni 4067-30* p. Ni 4067-2 4 p.Ni 'Strongest enhanced line of nickel. 4086-6 4081-5 fie- venal in dark H 4102-7 Ha 4101 -88 H 4102 Hy 4101-8 10 Ha 4173-1 He* 4174-5 p. Fe 4175 p.Fe 4173-52 j 4178-95 1 I'" F i 4173-5 | 4179-0 6-7 p. Fe 6-7 p. Fe JTe-A. 4169. 4185-1 ! 4200-6 p.H 4217-0 - -- 4235-2 4232-5 p. Fe 4232 p. Fe 4233-2", p. Fe 4233 3 8 p.Fe 4240-.; 4271-4 Neb.* 4267-2 4269 52 p. Cr 4269-8 1-2 p.Cr Nebular band A4265. 4283-7 4284-38 p. Cr 4284-4 2 p. Cr 4294 8 j 4294-20 p.Ti 4294-2 4 p. Ti (4296-72 p.Fe 4296-7 4 p. Fe i 4-2 5 o o o 35 o o PLATE IV. . 7 . G 1*87) .O 170 ' THE IILH INy.lV. . RELATION* OF NOV.E TO MILKY WAY. Return to desk from which borrowed Tins book is DUE on the last date stamped below. 21-100m-l V 49 (B7146sl6)476 YF SOLAR PHYSICS COMMITTEE- continued. THE SUN'S SPOTTED AREA. 1832-1900. A Statement of the Mean Daily Area in each Synodic Rotation of the Sun. Based upon Data collected at the Solar Physics Observatory. (1902.) Price Is. Qd., post free Is. 8d. SURFACE-AIR CIRCULATION. Southern Hemisphere Surface-Air Circulation. Being a Study of the Mean Monthly Pressure Amplitudes, the Tracks of the Anticyclones and Cyclones, and the Meteorological Records of several Antarctic Expeditions. (1910.) Price Qs., post free 6s. 4d. TE Spectroscopic c Sources, with speci 3s. 3d. Table of Wave 1 GREENWICH AS1 ATIONS, 1911. Or in Separat RESULTS. Do. A\rr> PFTFSTT\T. \^ UNIVERSITY OF CALIFORNIA LIBRARY -^OURCES. rial and Celestial Light- 07.) Price 3s., post free ES. post free Is. ROLOGICAL OBSERV- 5ree 5s. Qd. AIAGNETICAL AND METEOROLOGICAL. 1911. (1913.) Price 3s., post free 3s. 4d. Do. PHOTO-HELIOGRAPHIC. 1911. (1913.) Price 3s., post free 3s. 3d. REPORT OF THE ASTRONOMER ROYAL TO THE. BOARD OF VISITORS of the Royal Observ- atory Greenwich, June 7, 1913. (1913.) Price Is., post free Is. Id. CHRONOMETER WATCHES ON TRIAL for purchase by the Board of Admiralty. Rates of. Aug. 3 to Nov. 23, 1912. (1913.) Price 2d., "post free 2d. NOTE. The above are piMifhed yearly. GREENWICH OBSERVATIONS. APPENDICES. 1909. Investigation of the Motion of Halley's Comet from 1759 to 1910. (1912.) Price 3s., post free 3s. 4d. 1911. Altazimuth Catalogue for the Epoch 1900 '0. (1913.) Price 3s., post free 3s. 3d. CHRONOMETERS (BOX AND POCKET). ON TRIAL for purchase by the Board of Admiraltv. Rates of. June 17, 1911, to January 6, 1912. (1912.) Price Is., post free Is. SUN'S AXIS. The position of the Sun's Axis as determined from Photographs of the Sun, 1874-1912, measured at the Royal Observatory, Greenwich. (1913.) Price 2s. 6d., post free 2s. 9d. CAPE OF GOOD HOPE OBSERVATORY. ANNALS :- Vol. VIII. Part 3. Observations of Major Planets made with the Heliometer, 1905 to 1910. (1912.) Price 3s., post free 3s. 3d. Vol. X. Spectroscopic Researches : Part H. Appendix I. On the Spectrum and Radial Velocity of e Canis Majoris. (1912.) Price 6d., post free 7d. Part IV. On the Spectra of Graphites and " Lead Pencils," and on a convenient Comparison-spectrum. (1913.) Price Is., post free Is. lid. HISTORY AND DESCRIPTION OF THE OBSERVATORY. {1913.) Price 25s., post free 25s. 7d.