Types of Variable Stars

Algol variables (Beta Persei stars)

A subclass of eclipsing binary stars, named after Algol, where the brighter and more massive star is still on the main sequence while the less massive companion has evolved more and has become a subgiant.  This seemingly contradicts the theory of stellar evolution, which predicts that more massive stars evolve more rapidly, and is known as the Algol paradox.  It is explained (Crawford 1955) as a result of extensive mass transfer: the now less-massive star originally contained most of the system's mass, and it evolved rapidly beyond the main sequence.  As it expanded, this tar lost up to 85% of its mass to the companion (see: W Serpentis star) to end up as a faint low-mass sub-giant, while the companion became a massive hot and brilliant star, still on the main sequence.  Mass transfer continues at a very low rate in Algol systems, causing variations in the orbital period and feeble radio and x-ray emission.

As a result of the mass transfer, the two stars have the unusual property of being roughly the same size (several times larger than the sun) but having very different luminosities.  They thus have a light curve characterized by deep primary minima when the dim subgiant eclipses the bright main sequence star, alternating with scarcely detectable secondary minima when the subgiant is eclipsed.

Beta Cephei stars (Beta Canis Majoris stars)

A small group of pulsating variables, the prototypes being b Cep and b CMa.  They are hot massive luminous stars of spectral types O9 to B3 with short periods of rotation (about 3-7 hours) and a very small varation in visual brightness (0.01-0.25 magnitudes), although the range is much greater at ultraviolet wavelengths.  Some of the stars have two or even more periods of radial-velocity variation.  The pulsation mechanism is still uncertain.  Their position on the Hertzspring-Russell diagram for pulsating variables is just above the main sequence, beginning at type B3.

Cepheid Variables (W Virginis stars)

A large and important group of very luminous yellow giants or supergiants that are pulsating variables with periods ranging from about 1-70 days.  Over 700 are known in our galaxy and several thousand in the Local Group.  There are two categories: classical Cepheids (also known as Type I Cepheids) are massive young population I objects found in the spiral arms on the galactic plane; W Virginis stars (or Type II Cepheids) are much older and less massive population II objects found in the galactic centre and halo, especially in globular clusters, and are similar in distribution to the RR Lyrae stars.  The classical Cepheids are about 1.5-2 magnitudes more luminous than W Virginis stars of the same period.  The luminosity variations of both categories are continuous and extremely regular so that the periods can be measured very accurately.  Characteristic periods are 5-10 days (classical) and 12-30 days (W Virginis stars); the amplitudes are typically 0.5-1 in magnitude.  The prototype of the classical Cepheids is Delta Cephei, discovered in 1784.  The changes in brightness were found in the 1890s to be accompanied by and principally caused by changes in stellar temperature and also by changes in radius.  This was later explained in terms of pulsations in the outer layers of the stars that appear at a late evolutionary state (see pulsating variables).  The relation between period of pulsation and light variation of Cepheid variables was discovered during 1908-12 by Henrietta Leavitt.  Miss Leavitt could only measure the apparent rather than the absolute magnitude of the Cepheids under study.  An independent determination of the distance of a Cepheid of known period would lead to the graphical relation between period and luminosity, or absolute magnitude (see distance modulus).  It was quickly realized by Shapely that this period-luminosity relation was an invaluable tool for measurements of distance out tot he nearest galaxies and thus for studying the structure of our own galaxy and of the universe.  The Hubble Space Telescope will be able to study classical Cepheids in galaxies as far away as the Virgo cluster (about 15 megaparsecs).  Much work has been done to establish the graph of period versus absolute magnitude, mainly involving independent measures of the distances or luminosities of the Cepheids.  Baade and Kukarkin were consequently able to demonstrate in the 1950s the existence of the two distinct categories of Cepheids, having separate but parallel period-luminosity relations.

A period-colour relation was also discovered whereby the colour index of a Cepheid increases, i.e. becomes more positive, as the period increases: a relatively long period implies a redder and cooler star.  It was thus also possible to derive a period-spectrum relation between spectral type and period.  Both Cepheid categories follow these relations.

Delta Scuti stars

A class of relatively young pulsating variables, typified by Delta Scuti, that are about 1.5-3 solar masses and range in spectral type from about A3 to F6.  They were originally grouped with the dwarf Cepheids.  They are numerouse but tend to be inconspicuous because of their small variations in brightness, usually 0.01-0.4 magnitudes.  Like dwarf Cepheids they have very short periods ranging from about 0.8-5 hours.  A typical feature is a variation in the amplitude and shape of the light curve; irregularities in the light-curve contour result from the superimposition of two or more harmonic modes in which the stars is pulsating.

dwarf Cepheids (AI Velorum stars; RRs stars)

A group of pulsating variables that until recently were classed as RR Lyrae stars but have been shown to have a lower luminosity in addition to a shorter period of about 1.3 to 5 hours.  The variation in luminosity (i.e. amplitude) ranges from 0.2 or 0.3 up to 1.2 but can be higher.  At the lower amplitudes they are difficult to differentiate from Delta Scuti stars but are apparently older.  Instabilities of period and light-curve shape are characteristic of these stars and many exhibit the Blazhko effect found in some RR Lyrae stars.  The pulsations are probably mixtures of oscillaiton sof a fundamental mode and those at higher harmonic frequencies.

Long-period variables

A group of variable stars that are principally Mira stars but sometimes the red semiregular stars, type SRa, are included in the group.  The longest periods (up to 5 years) are found in the OH/IR stars, normally observable only at infrared wavelengths.

Mira stars (Mira Ceti variables)

Long-period pulsating variables, either red giants or red supergiants, that have periods ranging from about 80 to over 600 days and a range in brightness at about 2.5 magnitudes and sometimes exceeding 10 magnitudes.  They are numerous, Mira Ceti being the prototype.  The shape of the star's light curve is not constant, the maximum brightness varying quite considerably between periods (see Mira).  Because of the large amplitude they are easily recognizable, their high luminosity permitting detection at great distances.

Although their visual range is large, the range in bolometric and infrared magnitudes is very much less, may of them being infrared sources.  In terms of their spectra 90% can be classified as Me stars, the rest as either Ce (carbon stars)  or Se (zirconium stars), i.e. there are bright emission lines present in the spectra in addition to molecular bands.  The pulsations of these huge stars are not very stable.  There is evidence of shock waves developing within the tenuous atmosphere and traveling outwards, thus heating the gas and causing the production of emission liens.  The expanding envelopes often contain considerable dust grains, which produce detectable infrared emission and simple molecules.  Mira stars often show maser emission from hydroxyl water, and silicon monoxide molecules in the outer atmosphere.  See also OH/IR star.

Pulsating variables

Variable stars that periodically brighten and fade as a result of large-scale and more or less rhythmical motions of their outer layers.  The simplest motion is purely radial, a cycle of expansion and contraction in which the star remains spherical but changes in volume.  The idea that a periodic stellar expansion could lead to variable light output was proposed by Shapley in 1914, with Eddington presenting the theory in 1918.  The period of light variation is equal to the period of pulsation and is normally approximately constant.  The spectrum of the star also changes periodically due to changes in surface temperature.  The pulsation cycle is demonstrated by variations in radial velocity.  The pulsation period, P, was showwn by Eddington to be related to mean stellar density, r, by the period-density relation and hence to the stars radius, R, and Mass, M: P order of 1 / sqrt(r) order of sqrt(R³ / M) Since luminosity is also proportional to radius, pulsation period should be related to luminosity.  Types of pulsating variable include Cepheid variables, RR Lyrae stars, dwarf Cepheids, Delta Scuti stars, long-period Mira stars, semiregular variables (including RV Tauri stars), irregular variables, and Beta Cephei stars.

A star pulsates because there is a small imbalance between gravitational roces and outward directed pressure so that it is not in hydrostatic equilibrium.  When it  pulsates it expands past its equilibrium size untl the expansion is slowsed and reversed by gravity.  It then overshoots its equilibrium size again until the contraction is slowed and reversed by the increased gas pressure within the star.  The fact that the pulsations do not die away as energy is dissipated means that some process is convereting heat energy into mechanical energy to drive the pulsations.  The centre of the star is not involved in the pulsation.  For most pulsating variables (probably excluding Beta Cephei stars) the driving force is a valve mechanism in the form of a region of changing opacity near the stellar surface where the pulsation amplitude is greatest.  This region is an ionization zone in which atoms, mainly helium or hydrogen are partially ionized.  Normally atoms become transparent, letting heat escape, as they are compressed.  In an ionization zone the opacity increases with compression and the zone effectively acts as a heat engine: as long as it lies at the required depth below the surface it can drive the pulsations.  If the zones are too close to the surface, as in a very hot star, or if too deep, then the pulsation is complicated by the fact that the gas is thought to be able to undergo oscillations (as in an open musical pipe) either in a fundamental mode or as a harmonic of the fundamental.

The distribution of pulsating variables ont he Hertzspring-Russell diagram differs from that of normal stars.  Most lie on a nearly vertical band -- the Cepheid instability strip -- that extends upwards from the main sequence (and possibly below it) and merges into a broader instability region at the top right.  Stars reach these instability regions by various evolutionary paths and may turn up there several times during their evolution.  The existence of such instability regions indicates that with certain combinations of stellar luminosity and effective temperature a state of pulsation rather than rest is favoured.

RCrB Stars (R Coronae Borealis)

Stars that are extremely helium rich and growing hotter through gravitational collapse.  They are believed to be the result of collisions between two helium rich white dwarfs or a carbon-oxygen white dwarf and a helium white dwarf.  See Star Stuff Press Release.

RR Lyrae stars

A large group of pulsating variables that are very old giant stars (halo and disc population II stars) and are found principally in globular clusters.  They usually have periods of less than one day and median spectral types in the range A7-F5.  They were discovered in 1895, by Solon I.  Bailey, the group being named after RR Lyrae, discovered in 1899.  Available evidence indicates that all RR Lyrae stars have about the same mean absolute magnitude (about +0.5); they can therefore be used as distance indicators up to about 200 kiloparsecs (see distance modulus).

RR Lyrae stars are of a mixed nature.  They were divided by Baily into three groups -- RRa, RRb, and RRc -- depending on period and assymmetry of their light curves (Groups a and b are often combined today as RRab stars.  Other groupings have been made according to period.  Many RR Lyrae stars show a periodic variation in both period and shape of the light curve -- the Blazhko effect.  In addition the periods of some RR Lyrae stars are slowly changing at a constant rate as predicted by evolutionary theory, while others show abrupt changes in period.  The pulsations of these stars are very complex and can be subdivided into one group (RRab) oscillating in fundamental mode and a second group (RRc) oscillating in first harmonic mode.
 
 

Subgroups of RR Lyrae stars
type	approx. amplitude	period  (days)	light curve asymmetry
RRa	1.5-2	0.35-0.55	highest
RRb	1.2-7	0.5-0.8	intermediate
RRc	0.5	0.2-0.4	lowest

RV Tauri stars

A small group of very luminous pulsating variables, typified by RV Tauri, R Scuti and R Sagittae, that are principally G and K stars with some F stars.  They are yellow supergiants with extended atmospheres of gas that emit infrared radiation and have possibly been driven off by the pulsations.  They have very characteristic light curves with alternating deep and shallow minima and periods ranging from 20-145 days.  Since the luminosity fluctuations can be disturbed quite significantly in shape, period, etc., being most pronounced for longer-period variable stars, they are classified as semiregular variables.  RV Tauri stars can be distinguished from other similar yellow semiregular stars by the variation in their colour index, which mimics the light curve but goes through its maximum a little before the luminosity minimum.  A small group have double periodicity; DF Cygni has two separate luminosity oscillations, a rapid 50 day oscillation being RV Tauri stars superimposed on a much slower 780 day oscillation with a much greater amplitude.

semiregular variables  (SR Variables)

A heterogeneous group of giant and supergiant pulsating variables showing brightness variations that do not usually exceed one or two magnitudes, that have a noticeable periodicity ranging from several days to several years, but are also disturbed at times by various irregularities.  The light curves have diverse shapes.  Semiregular regulars have been subdivided into four subgrous: SRa and SRb variables are giants of late spectral type (M, C, and S) either with relatively stable periods (SRa) or ill-defined periods (SRb) and rare difficult to differentiate from long-period Mira variables; SRc variables are red supergiants of late spectral type, such as Betelgeuse, Antares, and Mu Cephei; SRd variables are highly lumonous yellow (F, G, and K) supergiants and giants.

T Tauri stars

Irregular variable stars of late spectral type whose spectra is dominated by strong emission lines, usually attributed to chromospheric activity and stellar winds in these stars.  They are invariably embedded in dense patches of gas and dust that may require observations in the infrared: the dust absorbs the visible light of the star and reradiates it at longer wavelengths.  The prototype is T Tauri.  The T Tauri stars are usually found together in groups (T associations) and are the youngest optically observable stages in the life of a star of about the sun's mass; more massive counterparts are observed as Ae and Be stars.  They are frequently associated with Herbig-Haro objects.

There is much evidence that T Tauri stars are young objects, for instance they have a high abundance of lithium, an element destroyed fairly early in a star's life, and they are surrounded by gas and dust.  They are thought to be young protostars that have only recently contracted out of the interstellar medium.  Lying above the main sequence in the Hertzspring-Russell diagram, they are still contracting and losing mass.  Their spectral lines reveal that some are extremely rapid rotators, throwing off material at speeds of up to 300 km/sec.  The irregular light variations are believed to arise partly from activity in the chromospheres of these young stars, and partly by the obscuring effect of the patchy dust in the coccoon as it moves in front of the star.

Wolf-Rayet stars (WR stars; W stars)

A small group of very luminous very hot stars, with temperatures possibly as high as 50,000 kelvin, that have anomalously strong and broad emission lines of ionized helium, carbon, oxygen, and nitrogen, but few absorption lines.  Since their discovery (by C. J. E. Wolf and G. Rayet, 1867), over 300 have been found in our Galaxy and its neighbors.  The majority either have lines of He, C, and O -- WC stars -- or He and N -- WN stars; both types have anomalously low abundances of hydrogen.  The emission lines are thought to arise in an expanding stellar atmosphere moving at very high speeds of up to 2000 km/sec. so that the star is continuously and rapidly losing mass.  The average mass for Wolf-Rayets is 10 solar masses.  About half are known to be binary stars, usually with O or B stars as companions, an example being Gamma Velorum (WC8 + O7).

These unusual properties provide clues that Wolf-Rayets are the centres of very massive Of stars stripped of their outer envelopes.  Since the Wolf-Rayets are the more evolved members of each binary, they must originally have been the more massive partner, a star of at least 20 solar masses.  Half that mass has thus been lost in their stellar winds, whose high outflow rate would strip this mass off the star in only 100,000 years.  This gas is in fact often visible as a ring nebula surrounding the Wolf-Rayet star.  In a close binary the companion's gravity may assist in the stripping, but for single stars the cause of the high mass loss is still uncertain.

W Serpentis star (Beta Lyrae star)

A close binary star system where matter is being transferred very rapidly from one star the other.  This occurs when the more massive member of a close binary evolves to become a red giant (the previous state may be an RS CVn star); the mass transfer can be so efficient that 85% of the red giant's mass is transferred to the other star.  The system ends up as an Algol variable.

Observationally, W. Ser stars are spectroscopic binaries characterized by emission lines from an extended gas envelope that is hotter than either of the stars in the system.  The emission is thousands of times brighter than the lines from the star's chromosphere.  This gas represents part of the giant star's mass that is lost to the system during the rapid transfer.  The bulk of the gas forms an accretion disc around the giant's companion and as this gas spirals inwards the inner regions can heat up to 100,000 K and supply the radiation that ionizes the extensive tenuous gas producing the emission lines.

The outer cooler regions of the accretion disc can camouflage the accreting star and make it look  larger and cooler than it actually is.  This star will often have the appearance of a giant rather than that of the underlying main sequence star.  When mass transfer is most rapid, as in Beta Lyrae, the disc conceals this component entirely.  Thus in Beta Lyrae we detect only the expanding giant star, a supergiant of spectral type B8.5 that is about 13 times the diameter of the sun and elongated towards its companion because it fills its Roche lobe (see equipotential surfaces).  The companion is hidden by a disc about twice as wide but only half as thick as the supergiants diameter.  The system is an eclipsing binary with disc and star alternately eclipsing one another; the ellipsoidal shape of the supergiant causes the light curve to peak between eclipses, when the maximum extent of the star is seen.  The B8.5 star is losing mass at a rate of about 10^-5 solar masses per year and this causes an increase in the orbital period (12.9 days) at a rate of 19 seconds per year.

W Ursae Majoris stars

A class of contact eclipsing binaries that have very short periods amounting to only a few hours and components that are so close they are grossly distorted by tital forces into ellipsoidal shapes.  Mass transfer occurs between the stars (see equipotential surfaces).  The components are approximately equal in brightness, as seen by the equal depths of the minima of the light curves.  As with W Serpintis stars, the light curves show continuous variation resulting from the distorted stellar shapes and variable surface intensity.  The two components are more similar in luminosity than in mass (the ratio of the masses can be 12:1), so energy from the more massive star's core is being fed into the companions photosphere.  The resulting chromospheric activity is thought to be responsible for the unusually high x-ray output of these stars.

About 0.1% of all main-sequence stars should be born in close enough doubles to become contact systems: the lower mass F, G, and K stars are the W UMa class while the rare more massive analogues are called SV Centauri stars.  In both cases as the more massive component swells to become a giant, its outer layers will surround both stars to create a common envelope star; the final outcome will be a coalesced star or a cataclysmic variable system.

Ae stars

Hot stars of spectral type A that in addition to the normal absorption spectrum have bright emission lines of hydrogen.  These lines are thought to arise in an expanding atmospheric shell of matter lost from the star.  Like some Be stars they are probably similar to T Tauri stars except that they have larger masses.

Am stars

Peculiar main-sequence stars of spectral types from A0 to F0 in which there is an over-abundance of heavier elements and rare earths and (less so) of iron, and an apparent under-abundance of calcium.  They rotate slower than normal A stars and are almost all short-period spectroscopic binaries.  Current theory suggests that tidal effects slow the A star's rotation, leading to an unusually stable atmosphere where heavy elements can diffuse up from the interior.

Ap stars

Peculiar main-sequence stars of spectral types from B5 to F5 in which spectral lines of certain elements (mainly Mn, Si, Eu, Cr, and Sr) are selectively enhanced and are somtimes of varying intensity (see spectrum variables).  The enhancement and its variation is apparently associated with strong and often variable magnetic fields.  Like the Am stars, Ap stars are generally slow rotators, but they differ in being single stars rather than spectroscopic binaries.  The relative enhancement may be due to diffusion in the stable atmosphere caused by slow rotation, modified by the effect of strong magnetism.  The spectral peculiarities could also relate to stellar temperature,  the manganese stars, being hotter than the silicon stars, which in turn are hotter than the europium-chromium-strontium stars.  See also magnetic stars.

Be stars

Irregular variables of spectral type B in which bright emission lines of hydrogen are superimposed on the normal absorption spectrum.  They are now known to be identical to shell stars.  Some Be stars are young stars, rather more massive than Ae stars: together they are sometimes classed as Herbig emission-line stars, and are heaver versions of T Tauri stars.  These Be stars are rotating very rapidly and are slowly losing mass to an expanding shell that surrounds the star and is drawn into a disc around the equator due to the rotation.  Other Be stars are in close binary systems, with shells that consist of gas being accreted from an evolved companion star (see mass transfer).

Rare red giant stars of low temperature that have an over-abundance of carbon relative to oxygen and also an unusually high abundance of lithium. WZ Cassiopeia is an example.  The spectra show strong bands of carbon compounds, including C₂, CN and CH.  In the earlier Harvard classification (see spectral types) these stars were divided into R stars and N stars: N stars are similar to M stars, being cooler and much redder than the K-type R stars.  See also S stars.

M Stars

Stars of spectral type M, cool red stars with surface temperatures (T_eff) less than about 3500 kelvin.  Molecular bands are prominent in the spectra, with titanium oxide (TiO) bands dominant by M5.  Lines of neutral metals are also present.  Some show emission lines too, and are classed as Me stars.  Antares and Betelgeuse are M stars.  See also carbon stars; S stars.

Me Stars

Cool red stars whose spectra show emission lines of hydrogen in addition to the molecular bands of titanium oxide characteristic of M stars.   This occurs in both giant and dwarf (main-sequence) M stars, apparently for different reasons.  The giant Me stars are mainly long period variables such as Mira stars, whose distended atmospheres are losing matter to space.  The emission lines from dwarf Me stars seem to be related to the presence of strong magnetic fields; they are often flare stars.

MS stars

 See S Stars.

Of stars

Young massive O stars belonging to earlier subdivisions than O5.  In addition to a well-developed absorption spectrum they show selectively enhanced emission lines of ionized helium (He II) and nitrogen (N III) that can vary in intensity in an irregular manner.  The emission lines arise in an unstable atmosphere that is being lost from the star.  Of stars are the hottest, most luminous, and probably the most massive stars in the Galaxy: the curreent record holder is HD 93129A, near Eta Carinae, with a temperature of 50,000 K, a luminosity of 3,000,000 suns and a mass of probably 120 solar masses.

OH/IR star

A very large extremely cool giant or supergiant star that is losing mass very rapidly and is detected only bi its infrared radiation and by its hydroxyl (OH) maser emission (see maser source). The first OH/IR stars were found in a survey of OH maser sources in 1973; many more were detected by IRAS's infrared survey in 1983.  These stars are usually very long period variables.  Mira stars often show similar but weaker maser and infrared emission.  The OH/IR stars, however are losing mass a hundred times faster than the Mira stars.  They are surrounded by a very dense shell of cosmic dust that shows absorption typical of silicates.  The shells have temperatures ranging from 1000 K to only 100 K.  Since the stars lose a solar mass of gas and dust in only 10,000 to 100,000 years, they must be changing very rapidly to their next evolutionary state, probably a planetary nebula.

RS Canum Venaticorum star (RS CVn star)

A short-period binary star system that contains a subgiant of spectral type G or K exhibiting intense activity at radio, ultraviolet, and X-ray wavelengths; the other component is a normal F or G main-sequence star.  The orbital period can range from 1 day to 2 weeks.  The two stars are not undergoing mass transfer (they are a detached system), but tidal forces have locked the rotation of the subgiant so that it is forced to rotate once in each orbital period, much faster than a normal subiant.  This rapid rotation apparently causes strong magnetic fields, which then produce the observed activity.

The light curve of these stars shows a continuously changing brightness during the orbital period, believed to be due to dark 'starspots' extensively blanketing one half of the subgiants's surface.  Measurements of the Zeeman effect in light from Lambda Andromedae show that the spotted hemisphere has a field strength of over 0.1 tesla, similar to that of sunspots.  Some of these stars, including the prototype RS CVn, are also eclipsing binaries; the light curve shows regular dips of about a magnitude as the subgiant eclipses the main-sequence star.  The subgiant's magnetic field also causes intense radio flares (compare flare stars), strong ultraviolet emission lines from the chromosphere, and powerful x-ray emission from a very hot corona.

As the subgiant swells to giant size, it will begin to transfer mass and become a W Serpentis star and then an Algol variable.

S stars (zirconium stars; heavy-metal stars)

Stars of spectral type S, about half of which are irregular long-period variables.  They are red giants similar to M stars but distinguished spectroscopically by the presence of molecular bands of the heavy-metal oxides of zirconium (ZrO), lanthanum, yttrium, and barium rather than the oxides of titanium (TiO), scandium, and vanadium.  Lines of technetium (longest half-life 2x10⁶ years) can also be present.  Pure S stars have very strong ZrO bands with either very week or no TiO bands.  MS stars are M-type stars showing ZrO bands.

Dwarf Novae

A small group of intrinsically faint stars that are characterized by sudden increases in brightness occurring at intervals of a few weeks or months, the maximum brightness lasting only a few days.  The change in brightness (i.e. amplitude) is between 2-6 magnitudes.  The first to be discovered, U Geminorum is typical of the majority, which are therefore classified as U Geninorum stars.  This subgroup displays a fairly smooth decline in brightness from the maximum, unlike the much smaller subgroup of Z Camelopardalis stars that can undergo standstills, i.e. periods of nearly constant intermediate brightness, before dropping to minimum brightness.  Both the occurence of the standstill and its duration - a few days to many months - are quite unpredictable.  There are also periods of erratic light variations.  The brightest and most carefully observed dwarf nova is SS Cygni.

Dwarf novae are a class of cataclysmic variables, i. e. close binary stars in which the primary is a white dwarf.  The secondary is a cooler main-sequence star, spectral type K or G.  The components have similar masses (about 0.7 to 1.2 solar masses); the orbital periods are between 3 to 15 hours.  The secondary is undergoing irregular expansion and in doing so can fill one Roche lobe of the equipotential surfaces of the system.  Hydrogen-rich gas then streams from the secondary and takes up a disc-shaped orbit around the primary.  The outbursts occur when this disk brightens up, probably as a result of instabilities when its hydrogen changes from opaque to transparent and back.  It does not involve an explosion, and no significant amount of mass is ejected.  The gas in the disc spirals down on to the white dwarf, where it may eventually cause a nova explosion.

Nova

A close binary star system in which there is a sudden and unpredictable increase in brightness by maybe 10 magnitudes (see Novas Table).  Novae are a class of cataclysmic variable.  In a typical spiral galaxy like our own, there may be 25 nova eruptions per year.  The brightness increases to a maximum within days or stometimes weeks and then declines to a value probably close to its faint pre-nova magnitude, indicating that the eruption did not disrupt the bulk of the star. Fast novae usually increase in brightness by a factor of 10⁵ in a few days, remaining at peak brightness for less than a week; they then decline steadily, initially quite rapidly, over serveral months. Slow novae reach maximum brightness more slowly and erratically, the increase being less than fast novae, and then decline much more slowly.  The total energy released however is about the same in both cases.  Hydrogen-rich gas is ejected from the star resulting in the tremendous outflow of heat and light.  The ejected matter forms a rapidly expanding shell of gas that can become visible as the nova fades.  At later stages of the eruption the spectra of most novae show the bright forbidden lines charactristic of very low density emission nebulae

Like other cataclysmic variables, a nova is a close binary system in which one component is a white dwarf.  The other is a main-sequence star that is expanding to fill its Roche lobe (see equipotential surfaces) and is hence losing mass to the white dwarf: some of its gases 'overflow' to form a disc surrounding the white dwarf.  The rate of outflow is about 10^-9 solar masses per year, about ten times higher than in a dwarf nova system, and as a result the disk is always as bright as a dwarf nova at maximum.  Before a nova explosion the hot turbulent disk is the brightest part of the system, and its light makes a pre-nova appear as a bluish irregularly fluctuating 'star'.  The hydrogen in the disk spirals down on to the surface of the white dwarf, and after a period of some 10,000 to 100,000 years enough has accumulated to react in a thermonuclear explosion -- the nova outburst.  The explosion leaves the system fundamentally unchanged, however, and the flow of gas resumes to reestablish the accretion disc around the white dwarf.

Novae are now named intially by constellation and year of observation; before 1925 they were numberd in order of observation.  They are later given a variable star designation, as for example with DQ Herculis, i.e. Herculis 1934.  See also recurrent nova; x-ray transients.

		mv
nova	type	before	max
Nova Persei 1901 (GK Per 2)	fast	13.5	+0.2
Nova Aquilae 1918 (V603 Aql 3)	fast	10.6	-1.1
Nova Pictoris 1925 (RR Pic)	slow	12.7	+1.2
Nova Herculis 1934 (DQ Her)	slow	14.3	+1.4
Nova Cygni 1975 (V1500 Cyg)	fast	> 20	+1.8

SS Cygni

The brightest and most carefully observed dwarf nova.  Normally of 12th magnitude, it rises to maybe 8th magnitude every month or so: its period varies widely from the mean of 51 days; the maxima vary in shape, brightness, and duration.  Like other dwarf novae it is a close binary system in which one component is a white dwarf; the other is a normal G5 star.  The orbital period is 6.5 hours.  SS Cygni is also an x-ray source, i.e. an x-ray binary.

Examples:

T Pyxidis (1890, 1902, 1920, 1944, 1965), RS Ophiuchi (1901, 1933, 1958, 1967), T Coronae Borealis (1866, 1946).

1. Robert J. Bradbury, "When Stars Go Dark" (2000)

1. The Facts on File Dictionary of Astronomy, 4th Ed., Valerie Illingworth, John O. E. Clark (Eds.)  (2000)
2. GCVS Types: P. N. Kholopov, General Catalog of Variable Stars, 4th Ed. (Moscow: Nauka) (1997).
• GCVS Description
3. Marcos J. Montes, "Supernova Taxonomy" (17 Oct 1996).
4. Marcos J. Montes, "Compendium of Supernova and Supernova Remnant Resources" (24 Apr 2001).
5. "Freak Star Mystery Solved", Star Stuff Press Release (27 May 2002)