Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://star.arm.ac.uk/~ambn/318jgd.ps Äàòà èçìåíåíèÿ: Wed Sep 8 13:27:21 1999 Äàòà èíäåêñèðîâàíèÿ: Mon Oct 1 21:07:13 2012 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: m 106

Mon. Not. R. Astron. Soc. 000, 000--000 (0000) Printed 7 August 1997 (MN L a T E X style file v1.4)
Evidence for an intra­binary radio emission region in the
Algol system V505 Sagitarii
A. G. Gunn 1;2 , V. Migenes 3? , J. G. Doyle 2 and R. E. Spencer 1
1 University of Manchester, Nuffield Radio Astronomy Laboratories, Jodrell Bank, Lower Withington, Cheshire, SK11 9DG, UK
2 Armagh Observatory, College Hill, Armagh, BT61 9DG, N. Ireland
3 ATNF, P. O. Box 76, Epping, NSW 2121, Australia
Accepted 1997 Received 1997 in original form August 1997
ABSTRACT
We present radio interferometric observations of the Algol­type binary system V505
Sagitarii made with the ATNF Compact Array at 6 cm and 3.6 cm over one orbital
cycle (1.18 days). We obtained a strong detection of the source (1.5 mJy at 6 cm and
1.4 mJy at 3.6 cm). The radio flux level shows a clear modulation with evidence of
eclipses of the emission region at both conjunctions of the binary which may indicate
the existence of an intra­binary region of activity. This has important consequences
for the details of coronal formation and field interaction in active close binary stars.
Key words: stars: activity ­ binaries: close ­ stars: flare ­ stars: late type ­ stars:
variables: other ­ radio continuum: stars
1 INTRODUCTION
The Algol­type binaries (McCluskey 1993; Batten 1981; Shu
& Lubow 1981) are a well­studied group of eclipsing systems
named after the prototype Algol (fi Persei). The classical
EA2 Algol binaries (Budding 1986) are semi­detached sys­
tems consisting of an early­type primary and a late­type
secondary. Photometric investigations reveal that the cooler
components of Algols appear to have spectral types, radii
and rotation rates similar to the active components of the
RS CVn binaries (Hall 1976). The main difference is that
the active stars in Algols are significantly evolved, invariably
contact their Roche surfaces and have lost a substantial frac­
tion of their mass. Since UV, IR and optical spectra of Algols
are dominated by their hotter components it is extremely
difficult to study the activity signatures of their cooler com­
ponents. However, because Algols are tidally locked like RS
CVns they also have high rotation rates and therefore also
display magnetic activity in the cooler stars. For example,
during the (total) primary eclipse of U Cep the hotter star is
totally obscured producing a deep nearly flat­bottomed pho­
tometric light curve. IUE spectra of this system (Gimenez et
al. 1990) reveal a rich emission line spectrum similar to those
arising from the chromospheres of RS CVn stars. U Cep has
also shown photometric evidence for cool photospheric star­
spots and flare­like events (Olson 1985; Richards 1990).
A great deal of evidence for solar­like activity in the cool
components of Algols can be found in the X­ray and radio
? Present Address: VSOP Project, National Astronomical Ob­
servatory, Osawa 2­21­1, Mitaka, Tokyo 181, Japan
regimes. Algol itself was first detected as an X­ray source
by Schnopper et al. (1976). A study of several systems by
White & Marshall (1983) indicated quiescent X­ray lumi­
nosities of 1--7 10 30 erg s \Gamma1 which are very similar to those
observed on RS CVns. Another similarity with RS CVns
is that the X­ray data can be successfully modelled with a
bimodal coronal temperature distribution (T1 ¸ 1--6 10 6 K
and T2 ¸ 2--4 10 7 K). Time­resolved observations of Algol
in X­rays by White et al. (1986) and Stern et al. (1990) us­
ing EXOSAT revealed that the cool star's corona was not
eclipsed by the B­star and so must extend out to at least
one stellar radius (– 3R fi ). These results are extremely in­
teresting because a similar lack of coronal X­ray eclipses are
found for RS CVns. Hence the dimensions of X­ray coro­
nae in close binary systems do not seem to be primarily
dependent on the relative activity levels of the components.
However, a comparison of the X­ray emission in RS CVn
and Algol systems by Singh, Drake & White (1996) showed
similar X­ray luminosity distributions although Algols were
less luminous in X­rays by a factor of ¸3 when compared to
RS CVn stars with the same orbital period.
Not a great deal of work has been done on Algol­type bi­
naries in the radio regime despite the fact that Algol was one
of the first detected stellar radio sources (Wade & Hjellm­
ing 1972). Many stars, including close binaries, in which the
dynamo generation of magnetic activity is expected to op­
erate, show evidence for large coronal regions similar to but
more energetic than those in the Sun. In general it is found
that emission associated with flare events arises in regions
smaller than or comparable to a stellar radius while quies­
cent emission arises from an extended halo whose dimensions
c
fl 0000 RAS

2 A. G. Gunn et al.
are comparable to the overall size of the binary system. Of­
ten both forms of emission are present and such a structure
is said to have a core­halo morphology. Clark, Kellerman &
Schaffer (1975) showed that the dimensions of Algol in the
radio are considerably greater than the size of an individual
star and that the emission implied a brightness temperature
of 4 10 8 K from a plasma magnetised with fields of B ¸ 10
G. Mutel et al. (1985) applied the core­halo model of radio
emission to 5 GHz observations of Algol. This model was
found to be successful for both RS CVns and Algols. For
Algol they found that the core component was smaller than
the size of each star and originated in a high­temperature,
highly magnetised plasma (Tb ¸ 10 10 K, B ¸ 100 G). The
halo component was circularly polarized and appeared to
originate from a region comparable in size to the binary sys­
tem with Tb ¸ 10 9 K and B ¸ 10 G. A complex model of
the Algol system using VLBI observations was presented by
Lestrade et al. (1988). They suggested that gyrosynchrotron
emission arises from an extended coronal region with an al­
most uniform electron density and magnetic field which de­
creases slowly with radial distance. They further proposed
that flare events were triggered by magnetic reconnection
at local sites within the corona producing inhomogeneous
self­absorbed synchrotron emission.
Direct evidence for the large size of radio coronae in RS
CVns was presented by Owen & Spangler (1977) and Doiron
& Mutel (1984) who both failed to detect a radio eclipse in
the eclipsing system AR Lac. VLBI observations by Mutel et
al. (1985) have also indicated the extensive nature of radio
coronae. These authors presented a time dependent model
of the emission to account for the core­halo morphology in
which optically thick compact regions on the surface of the
cool star expand as coronal loops, eventually becoming ex­
tended optically thin structures. Such a model is consistent
with the evolution of magnetic loops in the Sun.
Another possibility for the large coronae of active bina­
ries is that of a joint magnetosphere between the two stellar
components. This can form since in RS CVns both stars
are expected to develop dynamo­generated magnetic fields.
Uchida & Sakurai (1983) suggested that the existence of stel­
lar differential rotation can lead to twisting of magnetic flux
tubes and the subsequent reconnection of loops can give rise
to the extended coronae. A similar proposition of interacting
magnetic flux tubes was used to explain the UV properties
of the RS CVn system UX Ari by Simon, Linsky & Schiffer
(1980). Currently this model suffers from the observation
of similar large radio emission regions in some active single
stars (Turner 1985) and Algol­type systems. However, re­
cent observations by Gunn et al. (1997) at 6 cm and 3.6 cm
provided evidence of an intra­binary region of activity in the
RS CVn star CF Tuc. Lestrade (1996) recently showed that
the preferred site of the radio emission on the RS CVn stars
UX Ari and oe 2 CrB is in the intra binary region. These re­
sults have important consequences for the details of coronal
formation and field interaction in active close binary stars.
Umana, Catalano & Rodono (1991) and Umana et al.
(1993) interpret three­frequency spectra of a small sample of
Algol binaries with a core­halo model. They found that for
all targets the observed spectrum could not be explained by
homogeneous sources with constant magnetic field strength
and uniform electron density for any electron energy distri­
bution they considered. They concluded that the emission
process was non­thermal in origin and showed that the core
components were smaller than the size of the active stars
with a magnetic field of 60--100 G. They proposed that a
search for modulation of the emission with orbital phase
would better locate these components. Not only do these
results concerning the dimensions of the radio emitting re­
gions in Algols correspond very well to those observed in
RS CVns but the levels of emission and the source detection
rates are strikingly similar. The average radio luminosity in
the Algol survey of Umana, Catalano & Rodono (1991) was
1.6 10 16 erg s \Gamma1 Hz \Gamma1 while the median value for RS CVns
in the survey of Morris & Mutel (1988) was 2.5 10 16 erg s \Gamma1
Hz \Gamma1 . Furthermore brightness temperatures of Tb ¸ 10 8 --10 9
K and magnetic fields of 10--100 G seem to be common for
both groups.
A comparison of the radio properties of RS CVn bina­
ries and Algol systems provides a viable test of the mech­
anism originally proposed by Uchida & Sakurai (1983) to
explain the intense and extensive emission from RS CVns.
In Algols no such mechanism should be possible but they ap­
pear to display quantitatively similar behaviour to the RS
CVns. What is unclear is whether the observed properties
of such binary­scale emission is directly influenced by the
proximity of the stars and/or their relative degrees of ac­
tivity. In this paper we present and discuss interferometer
observations obtained for the Algol binary V505 Sagitarii si­
multaneously at 6 cm and 3.6 cm over the orbital period of
the system. We present radio images at both frequencies for
this binary system and discuss the variation of the observed
flux levels.
2 V505 SAGITARII
V505 Sagitarii is a classical eclipsing Algol system with a
period of 1.183 days consisting of an A2V primary and a
G5IV secondary of radii 2.2R fi and 1.15R fi , respectively, at
a separation of about 7.0R fi . The secondary is believed to be
filling its Roche lobe and a tertiary component that orbits
the eclipsing pair has been detected. A preliminary solu­
tion for the orbital elements was provided by Slonim (1934)
(who found i = 82.5 ffi ) and the ephemeris was subsequently
improved by Lause (1935, 1938). Popper (1949) provided
a spectroscopic radial velocity curve but was unfortunately
only able to determine the elements of the primary com­
ponent. V505 Sgr has a long history of photometric obser­
vations. Times of minima and ephemeris calculations have
been provided, amongst others, by Chambliss (1972), Wal­
ter (1981a,b), Khalesseh & Hill (1991), Rovithis­Livaniou &
Rovithis (1992, 1994), Walker (1993) and Chambliss et al.
(1993). A spectroscopic study of V505 Sgr was performed
by Tomkin (1992).
V505 Sgr has not been extensively studied in the ra­
dio regime. It was included in the sample of active binaries
observed with the Parkes 64­m radio telescope by Slee et
al. (1987). They measured a flux density of 11.2 mJy at 8.4
GHz. During a VLA survey of Algol systems Umana, Cata­
lano & Rodono (1991) found a radio luminosity of 5.25 10 16
erg s \Gamma1 Hz \Gamma1 and a brightness temperature of 2 10 9 K sug­
gesting the possibility of a non­thermal emission mechanism.
More recently Umana et al. (1993) presented the radio spec­
trum for V505 Sgr at three frequencies and found it to be
c
fl 0000 RAS, MNRAS 000, 000--000

Evidence for an intra­binary radio emission region in the Algol system V505 Sagitarii 3
V505SGR IPOL 4800.000 MHZ IF1SUB8.ICLN.1
PLot file version 1 created 24­AUG­1995 21:53:49
Peak flux = 1.5276E­03 JY/BEAM
Levs = 4.7000E­05 * ( ­3.00, 3.000, 6.000,
12.00, 24.00, 48.00, 96.00)
DECLINATION
(J2000)
RIGHT ASCENSION (J2000)
19 53 07.2 07.0 06.8 06.6 06.4 06.2 06.0 05.8 05.6
­14 35 00
30
36 00
30
37 00
30
Figure 1. The ATCA radio map of the area around V505 Sgr at 6 cm. The restoring beam is shown in the upper left hand corner.
Plotted flux levels are at ­3, 3, 6, 12, 24, 48 and 96 times the RMS noise level.
characterised by a strong high­frequency component with a
much lower flux density at low frequencies. They discuss the
possible physical conditions in the radio emitting regions of
V505 Sgr and conclude that it is either one order of magni­
tude larger than the binary system or originates from large
magnetic structures in the outer coronal regions. V505 Sgr
does not appear to have been observed in the X­ray or UV re­
gions although Arevalo, Antonopoulou & Lazaro (1993) did
observe it in the infrared. Some general discussion on V505
Sgr is given by Chambliss (1992) and Giurkin, Mardirossian
& Mezzetti (1983).
3 OBSERVATIONS AND DATA REDUCTION
Observations were carried out with the 6­km Australia Tele­
scope Compact Array (ATCA) located near Narrabri NSW
and operated by the Australia Telescope National Facility
(ATNF) for CSIRO. The ATCA consists of five 22­m anten­
nae on a 3­km track together with an antenna located 6 km
to the west. A recent description of the array can be found
in Manchester (1991). Observations began at 18:19 UT on
11th April 1995 and consisted of six 7­hour segments of data
observed at the same LST each day. The diurnal variations
in quiescent flux level was assumed to be negligible so that
the flux variation with orbital period could be examined. A
total of 42 hours of data were obtained which was adequate
to cover the orbital cycle of 28.4 hours although during suc­
cessive epochs. The array was calibrated in amplitude by
convenient observations of the primary flux calibrator 1934­
638 with fluxes of 6.33 Jy at 6 cm and 2.59 Jy at 3.6 cm.
Phase calibration was achieved by observations of the com­
pact source 1937­101 taken every 15­20 minutes. The target
data consists of approximately 15 minute scans of 15 sec­
onds integrated amplitude and phase measurements. All ob­
servations were conducted at two simultaneous frequencies
of 4.8 GHz (6 cm) and 8.64 GHz (3.6 cm) with instrumen­
tal bandwidths of 128 MHz. The effective bandwidth was
approximately 84 MHz across twenty channels which were
then averaged.
The data were processed using the software packages
aips and miriad. Imaging of the target field revealed four
point­like confusing sources with flux levels significantly
above the noise. These were removed by fitting single 2D­
gaussian profiles in the image plane and subtracting these
models from the visibility data. Analysis of the final data set
involved vector averaging the visibilities (taken at the phase
center) and plotting these against time. To achieve this the
c
fl 0000 RAS, MNRAS 000, 000--000

4 A. G. Gunn et al.
V505SGR IPOL 8640.000 MHZ IF2SUB8.ICLN.1
PLot file version 1 created 24­AUG­1995 22:21:34
Peak flux = 1.3749E­03 JY/BEAM
Levs = 3.8000E­05 * ( ­3.00, 3.000, 6.000,
12.00, 24.00, 48.00, 96.00)
DECLINATION
(J2000)
RIGHT ASCENSION (J2000)
19 53 07.2 07.0 06.8 06.6 06.4 06.2 06.0 05.8 05.6
­14 35 00
30
36 00
30
37 00
30
Figure 2. The ATCA radio map of the area around V505 Sgr at 3.6 cm. The restoring beam is shown in the upper left hand corner.
Plotted flux levels are at ­3, 3, 6, 12, 24, 48 and 96 times the RMS noise level.
target position was accurately determined by fitting a 2D­
gaussian model and then shifting the data to align the peak
position with the phase center. This allows the mapping pro­
cedure to be by­passed and the data sampled with the full
temporal resolution of the array. However we elected to aver­
age the visibilities over several scan lengths. The penalty of
directly plotting visibilities is a poorer signal­to­noise ratio
and the different temporal weighting gives rise to small dis­
crepancies between the map and visibility determined mean
fluxes.
4 RESULTS
4.1 Radio images
Figures 1 and 2 show the restored images of V505 Sagitarii
at 6 cm and 3.6 cm respectively. The aips task mx was used
to obtain maps 512 \Theta 512 pixels in size. For these observa­
tions the synthesized beam had unequal dimensions in the
RA and Dec. directions giving differing resolutions in each.
The images were produced using RA and Dec. pixel sizes of
0.4 and 2.6 arcsec respectively at 6 cm and 0.3 and 2.0 arcsec
respectively at 3.6 cm. The resulting images show approxi­
mately circular unresolved sources at both frequencies with
differing axis scales. Small extensions of the lowest contour
levels are not thought to be real.
Figure 1 shows an unresolved point source with a peak
flux level of 1.53 mJy/beam. The angular extent of this
source (! 1.4 \Theta ! 8.0 arcsec) is comparable to the size
of the synthesized beam (¸ 2.0 \Theta 8.0 arcsec). The mean
J2000 position (ff=+19 h 53 m 06 s .39, ffi=--14 ffi 36 0 11: 00 40) is
consistent with the most accurate optical position allowing
for proper motion and the small difference between the op­
tical and radio reference frames. The measured RMS noise
level (5oe) in the image is 2 10 \Gamma4 Jy/beam. Also in the field
(but not displayed in Figure 1) were four unresolved objects
of peak flux levels 3.6 mJy, 0.6 mJy, 0.4 mJy and 0.4 mJy
at positions ff=19 h 53 m 00 s , ffi=--14 ffi 29 0 10 00 , ff=19 h 53 m
06 s , ffi=--14 ffi 40 0 05 00 , ff=19 h 53 m 11 s , ffi=--14 ffi 34 0 53 00 and
ff=19 h 53 m 05 s , ffi=--14 ffi 41 0 33 00 , respectively. The strongest
of these sources was used as a control source to check the
flux density variations of V505 Sgr and to give estimates of
random errors.
Figure 2 also shows an unresolved point source with a
peak flux level of 1.37 mJy/beam with an angular size of !
0.8 \Theta ! 5.0 arcsec, comparable to the size of the synthe­
sized beam (¸ 1.0 \Theta 5.0 arcsec). This detection is almost
c
fl 0000 RAS, MNRAS 000, 000--000

Evidence for an intra­binary radio emission region in the Algol system V505 Sagitarii 5
1
0.5
0
­0.5
0 0.5 1 1.5 2
0
2
4
Orbital Phase
Figure 3. (a) The V­magnitude light curve of V505 sgr observed by Chambliss et al. (1993). (b) Time evolution of the 6­cm and 3.6­cm
fluxes from V505 sgr folded over the orbital period and plotted against orbital phase. Also plotted are estimated error bars for the radio
data. Note that the data have been plotted twice to emphasise phase­correlated variations.
exactly coincident with the 6 cm image with a J2000 position
of ff=+19 h 53 m 06 s .39, ffi=--14 ffi 36 0 11: 00 50. The RMS noise
level (5oe) in this image was measured at approximately 2
10 \Gamma4 Jy/beam. Only one of the confusing sources found in
the 6­cm image was detected above the noise at 3.6 cm and
so could be used as an indicator of flux error levels. To aid in
comparison Figures 1 and 2 are plotted over approximately
the same coordinate ranges and with the same contour levels
as multiples of the RMS noise in each image.
An estimate of the random error in the flux levels was
made by analysing the strongest confusing source in the im­
age planes at both 6 cm and 3.6 cm for which we assume
there is no intrinsic variation. RMS values of 0.2 mJy and
0.3 mJy were found at 6 cm and 3.6 cm respectively. These
estimates compare well with the 5oe levels in the map (both
approximately 0.2 mJy). The mean 6 cm flux level derived
from the visibilities (1.5 mJy) compares well with the peak
flux in the image shown in Figure 1. The mean 3.6 cm flux
level of 1.4 mJy is also in good agreement with the peak flux
in the image shown in Figure 2. Based on the results of our
error estimates we believe that no flux variation less than
about 13% at 6 cm and less than about 23% at 3.6 cm can
be regarded as statistically significant.
4.2 Radio light curve
Figure 3 (b) shows the amplitude variation of the target
against orbital phase for both the 6 cm and 3.6 cm data.
The orbital phase was calculated using the photometric
ephemeris of Rovithis­Livaniou & Rovithis (1994). The es­
timated error in phase using this ephemeris is 0.008. Also
shown in Figure 3 (a) is the V­magnitude light curve of the
system plotted with data from Chambliss et al. (1993). Note
that the data in Figure 4 has been plotted twice (i.e. over
two orbital periods) to emphasize the variation of flux with
phase.
We see a clear modulation of the radio flux of V505
Sagitarii at both 6 cm and 3.6 cm. The flux levels appear
to decrease at phases centred approximately at 0.5 and 1.0
(i.e. at both conjunctions of the orbit). The decrease in flux is
about 75% of the mean continuum level for both wavelengths
and for both primary and secondary eclipses. Based on the
error estimates this is approximately 6 times the lower limit
of detectable variation at 6 cm and 3 times the lower limit
at 3.6 cm. The events are also above the 3oe level in the over­
all flux variations and so are physically significant. The flux
decrease at primary conjunction (phase 1.0) is broader than
c
fl 0000 RAS, MNRAS 000, 000--000

6 A. G. Gunn et al.
that at secondary conjunction (phase 0.5) and appears to
be centred symmetrically around the conjunction. The nar­
rower secondary flux decrease appears to be shifted slightly
to phases greater than 0.5. The light curve implies that ei­
ther a double coronal structure exists or a significant frac­
tion of the emission originates from a region between the
two stars. To test these hypotheses a relatively simple set
of geometric models were used to replicate approximations
to the light curve. All of these models assumed that the
emission regions are spherically symmetric and emit as ho­
mogeneous and isotropic sources. More complex geometric
models are not warranted for a comparison to data with
inherent stochastic noise. The best model light curve repli­
cates two unequal eclipses at conjunctions which have non­
zero flux and indicates that the secondary component has
a large coronal radius and produces a significant proportion
of the flux but some additional emission originating between
the two stars is required. This indicates that an additional
source of the radio emission in V505 Sgr is a structure that is
eclipsed both by the primary and the secondary component.
Our subsequent interpretation assumes that these flux vari­
ations are due to eclipsing behaviour and not to any other
stochastic flux variations such as flares.
4.3 Characteristics of the radio emission
In this section we derive some characteristics of the radio
emission under the assumption that it is due to gyrosyn­
chrotron emission from mildly relativistic electrons. This is
generally the emission mechanism invoked to explain the
quiescent microwave fluxes from active late­type stars. For
a discussion of emission mechanisms the reader is referred
to Drake, Simon & Linsky (1992) and Gunn et al. (1997).
A valuable diagnostic of the emission process is the
brightness temperature Tb ; the temperature of a black­body
which produces the same specific intensity as the source. At
6 cm Tb is given by
Tb = 7:1 \Theta 10 7 S6=` 2 ; (1)
where S6 is the 6­cm flux density measured in mJy and ` is
the source angular dimension in milli­arcseconds. An upper
limit to Tb can be inferred using the size of the intra­binary
region (2.7R fi ) since this dimension is clearly suggested by
the eclipsing light curve. Using a mean 6­cm flux for V505
Sagitarii of 1.5 mJy this gives an upper limit of 2.7 10 9 K.
Note that if the source size is equal to the size of the active
component Tb ¸ 1.6 10 10 K and if equal to the overall size of
the system is one order of magnitude less than this. Hence
even assuming a large projected area of the emission region
in V505 Sgr the brightness temperature is very high.
For gyrosynchrotron emission the frequency at which
the emission becomes optically thin, špeak , depends strongly
on the magnetic field strength and the average electron
energy. For a thermal population of electrons the flux Sš
should decrease as š \Gamma8 for š ? špeak rather than as š ff where
­0.5 Ÿ ff Ÿ ­5.0 for a non­thermal electron energy spectrum
with 2 Ÿ ffi Ÿ 7 (where N(E) / E \Gammaffi ). The dual frequency
observations of V505 Sgr at 4.8 GHz and 8.6 GHz have re­
vealed mean fluxes of ¸1.5 mJy and ¸1.4 mJy respectively.
The resulting spectral index of ¸ --0.2 is within the range
normally found for quiescent active star emission (Lestrade
1988). This is consistent with the flat high­frequency spec­
trum observed for this system by Umana et al. (1993) and
implies that Ü ¸ 1 over most of the source. Since the spec­
trum is essentially flat the brightness temperature Tb is de­
termined by the average energy of electrons. This energy (in
MeV) is given by
Eeff = 8:6 \Theta 10 \Gamma10 Tb ; (2)
which assuming Tb ¸ 2.7 10 9 K is 0.23 MeV. For gyrosyn­
chrotron emission with electrons at this energy the results
of Dulk (1985) indicate that the most important emission
harmonics lie in the range s = 10­100. Here s = š/šB where
šB is the gyromagnetic frequency and š is the observing
frequency. In this case the magnetic field is given by
B = š
2:8s ; (3)
where B is in Gauss and š is in MHz. For the 6­cm data
then the implied range of the magnetic field in the emission
region is 17 ! B ! 170 Gauss. This field is valid only for
the surface or regions of the source where Ü – 1 and can be
associated with the intra­binary component. Assuming for
the moment that the active star is the source of the magnetic
field and that this has a dipolar topology given by
B = Bo
i R
R?
j \Gamma3
; (4)
where Bo is the surface induction, R? is the stellar radius
and R is the radial position of the Ü ¸ 1 surface from which
it is assumed most of the emission originates, then taking R
as the radius of the midpoint of the intra­binary region (R ¸
3.6R fi ) the surface field is of the order 71 ! Bo ! 710. If the
field varies as an inverse square law then the surface fields
are lower (44 ! Bo ! 440). These surface fields, presumably
associated with the active component, are somewhat lower
than those derived for other Algol systems by Stewart et al.
(1989) but are similar to the spatially averaged plage mag­
netic fields observed in the Sun (Zirin 1988). One problem
with these deductions is that the field topology is unlikely
to be dipolar in nature.
It is possible to further infer the electron density in the
emission region as follows. Again assuming an optically thick
source and assuming the emission is gyrosynchrotron from
mildly relativistic electrons with a power index of ffi ¸ 3
then following Dulk & Marsh (1982) the peak in the radio
emission is given by;
špeak ú 1:5 \Theta 10 4 (NeL) 0:23 B 0:77 (5)
where a nominal viewing angle of ` = 45 ffi is assumed, L is
the length scale of the emission region, B is the magnetic
field and Ne is the electron density. Assuming špeak ¸ š (the
frequency of observation), B ¸ 17 G and L is equal to the
size of the intra­binary region then Ne ¸ 3.7 10 8 cm \Gamma3 . Since
the emission region is constrained by the binary separation
this is a reasonable lower limit and is similar to densities
reported in the halo components of RS CVn systems (Mutel
et al. 1984; Lestrade et al. 1988). These calculations show
that the emission from V505 Sgr is consistent with an intra­
binary region radiating by a gyrosynchrotron process from
mildly relativistic non­thermal electrons with inferred field
strengths and electron densities comparable with those seen
in other active binary stars.
c
fl 0000 RAS, MNRAS 000, 000--000

Evidence for an intra­binary radio emission region in the Algol system V505 Sagitarii 7
5 DISCUSSION
5.1 Comparison with previous observations
Umana et al. (1993) observed V505 Sagitarii with the VLA
at frequencies of 1.5, 8.4 and 15 GHz and found a spectrum
characterised by a strong high frequency component and
a much lower flux density at low frequencies. They found
that for all the Algol­type binaries in their sample the ob­
served spectrum could not be explained by homogeneous
sources with constant magnetic field strength and uniform
electron density for any form of electron energy distribution
they considered. They therefore proposed a core­halo mor­
phology. Such a two­component model was derived for Algol
itself and the RS CVn binary UX Ari by Mutel et al. (1985)
and Lestrade et al. (1988) using VLBI measurements. Those
results showed that the emission core was smaller than an
individual stellar diameter and the halo was comparable in
size to the binary system. The model implies that compact
optically thick regions on the surface of the cool star associ­
ated with active regions expand as coronal loops, eventually
becoming optically thin structures about as large as the bi­
nary system. However, based on the shape of the spectrum
for V505 Sgr, Umana et al. (1993) suggested the emission
was of a thermal origin localized in the extended corona
around the active star with typical temperatures of T ¸ 10 6
K. Using the optically thin part of the spectrum they calcu­
lated a coronal size of 1.9 10 13 cm and an electron density of
Ne ¸ 5.9 10 7 cm \Gamma3 . Assuming a lower coronal temperature
of 10 4 K the corresponding numbers are 4 10 14 cm and 2.4
10 5 cm \Gamma3 , respectively. Even with the higher coronal tem­
perature this implies the emission region is more than 10
times the size of the binary system and it is unlikely that
such high temperatures and densities could be maintained in
a region this size. It should be noted that the radio luminos­
ity measured by Umana et al. (1993) is not consistent with a
thermal emission mechanism and this result was based sim­
ply on the shape of the spectrum. They further assumed that
the thermal emission was coupled to the typical coronal tem­
perature of 10 6 K deduced for RS CVn systems using X­ray
observations (e.g. Swank et al. 1981). Umana et al. (1993)
do note however that the thermal­like spectrum of V505 Sgr
could be due to a core­halo structure if the compact region
dominates that of the halo. They calculate a halo size of 7
10 11 cm (10R fi ) with Ne ¸ 1.5 10 5 cm \Gamma3 and average mag­
netic field of B ¸ 10 G. For the core component the size
was 1.8 10 11 cm (2.6R fi ), Ne ¸ 2 10 8 cm \Gamma3 and B ¸ 40 G.
However based on the large size of the core component and
its low average B they concluded that the radio emission
may originate from outer coronal regions involving larger
magnetic structures than seen in other Algol systems. Note
that not only is the core size derived by Umana et al. (1993)
almost equal to the size of the intra­binary region but their
values for magnetic fields and electron densities are very
similar to those computed above using the gyrosynchrotron
assumption. It is therefore possible that Umana et al. (1993)
observed the same feature with the VLA but that its spa­
tial location could not be deduced from the observations.
The present observations favour an intra­binary site for the
emission but the question remains whether the existence of
an intra­binary emission region can be reconciled with other
radio observations of Algol and RS CVn systems.
5.2 The origin of intra­binary emission
In a previous paper (Gunn et al. 1997) we presented evidence
of an intra­binary radio emission region in the RS CVn star
CF Tuc and commented on the possible importance of field
interaction in these systems. We have further found evidence
of such a region in the V505 Sgr system which should not be
capable of such field interaction. Numerous models may be
proposed for the source of this emission in the intra­binary
region of active close binaries. The essential requirement is
that acceleration of the electron population in this region is
achieved with energies of 0.23 MeV. It may first be postu­
lated that the radio source is associated with a coincidental
region of increased activity which has formed preferentially
between the two stars. This would of course show regular
eclipses at both conjunctions but why such a region should
be favoured without being associated with binarity is not
clear. In this section we explore some of the possibilities for
the origin of intra­binary emission.
5.2.1 Interacting fields
It has been postulated that radio emission in RS CVn stars
is associated with a volume of substantial reconnection of
magnetic field lines from individual magnetospheres around
both components of the system (Uchida & Sakurai 1983).
Evidence is growing of the existence of the subsequent
intra­binary active regions in these stars (Doyle, van den
Oord & Byrne 1989; Simon, Linsky & Schiffer 1980; Gunn
et al. 1997; Lestrade 1996). This mechanism, however, seems
unlikely for Algol systems since the early­type primaries
have a very inefficient dynamo mechanism and hence a weak
magnetic field. Landstreet (1982) has demonstrated the lack
of measurable magnetic fields in a sample of 22 stars of spec­
tral types A0­F5. However, Lestrade et al. (1988) noted
that the dependence of the magnetic field in Algol was
milder than expected for a simple static dipole and con­
cluded that a joint magnetosphere may be present. This pos­
sibility certainly requires investigation since Stewart et al.
(1989) reported apparently magnetic related radio emission
from EA1­type Algol systems, that is close binary systems
in which both components are early­type stars. An interact­
ing field mechanism is still possible if it is postulated that
the A2V star has an intrinsic magnetic field which is not dy­
namo generated, similar to the Ap stars which contain fields
up to 20000 G (Borra, Landstreet & Mestel 1982). Obser­
vations at 6 cm by Drake et al. (1987) of magnetic Ap and
Bp stars showed that the emission was similar to quiescent
coronal emission from solar type stars but the field strengths
were generally an order of magnitude greater. The lack of
kilogauss magnetic fields in V505 Sgr and the apparently
normal optical spectrum (Khalesseh & Hill 1991; Tomkin
1992) indicate that an intrinsic magnetic field in the primary
is not the source of the particle acceleration. Furthermore, as
pointed out by Kitamura (1980), only one eclipsing binary
star is known to contain an Ap component (AR Aur). It is
possible that some A2­F0 components of eclipsing binaries
are chemically peculiar Am stars (Abt & Bidelman 1969)
but none are known to possess magnetic fields. It should be
noted that Stewart et al. (1989) used radio data to derive
surface magnetic fields comparable to those in Ap stars for
a small sample of Algols. However these values were based
c
fl 0000 RAS, MNRAS 000, 000--000

8 A. G. Gunn et al.
on single­dish observations at 8.4 GHz and may be signifi­
cantly affected by noise. In fact the maximum flux for V505
Sgr reported in this survey (Slee et al. 1987) is almost an or­
der of magnitude higher than that reported here so that the
derived magnetic fields are also likely to be over­estimated.
We believe that a magnetic field associated with the primary
component in V505 Sgr is unlikely to be the origin of the
intra­binary emission. However, the concept of a magnetic
field in the primary components of Algol binaries and indeed
in apparently normal early­type stars in general cannot be
completely ruled out by these observations.
5.2.2 Stellar wind
Radio emission has been observed in highly luminous stars;
generally early­type B and OB giants, Wolf­Rayet stars and
some M­type super­giants. These observations are usually
associated with stellar winds and variable mass loss rates
(Abbott 1985). For such stars in binary systems particle ac­
celeration as well as magnetic field amplification probably
takes place at the interface of the stellar winds or due to
the interaction of a wind with the magnetosphere of the
companion. Such objects have been shown to display pri­
marily thermal emission or sometimes non­thermal emission
derived from shock ionization. This type of process can be
ruled out for V505 Sgr since the primary component is not
luminous enough to give an appreciable stellar wind. How­
ever the radial dependence of magnetic fields in Algol ob­
served by Lestrade et al. (1988) caused them to speculate
on the possibility of a stellar wind mechanism.
5.2.3 Accretion stream
Another possibility that must be considered is whether
the radio emission is directly associated with an accretion
stream in V505 Sgr. According to the models of Lubow &
Shu (1975) V505 Sgr, which has a mass ratio of 1.9, would
have an accretion stream which leaves the secondary star at
the inner Lagrangian point at an angle of `s ! 20 ffi . For V505
Sgr the maximum attainable accretion stream angle beyond
which there is no impact region on the primary (and an ac­
cretion disk forms) is 24.2 ffi . Hence, on the basis of ballistic
calculations V505 Sgr is likely to have both an accretion
stream and an impact region. It should be noted immedi­
ately that the classical noisy photometric signature associ­
ated with accretion streams has not been observed for this
system. There is also no spectroscopic evidence for such a
feature although the system is not well studied spectroscop­
ically.
Using the expression of Pringle (1985) for the cross­
sectional area of an accretion stream then the maximum
projected area (in cm 2 ) of the stream is given by
As = 3 \Theta 10 6 T \Gamma 1
2
4 Rs
r
R 3
MG ; (6)
where Rs is the length of the stream before it impacts the
primary, R and M are the radius and mass of the secondary
component respectively, T is the photospheric temperature
in units of 10 4 K and G is the gravitational constant. As­
suming `s ¸ 20 ffi for V505 Sgr then we find Rs = 2.03 10 11
cm and As = 4 10 21 cm 2 . Following Blondin, Richards &
Malinowski (1995) this calculation has assumed the temper­
ature of the stream to be comparable to the photospheric
temperature of 5700 K and the secondary mass is taken as
1.15M fi . At the assumed distance of the system (128 pc)
this corresponds to a projected solid
angle\Omega of 2.5 10 \Gamma21
sterad. At the assumed temperature and with typical stream
densities of ne ! 10 10 cm \Gamma3 such a source would be optically
thick at 6 cm and emit thermally. The maximum expected
flux at 6 cm for V505 Sgr is therefore ¸ 1 10 \Gamma8 mJy. Hence
the radio emission cannot be originating in the thermal ac­
cretion stream in this system. The effect of any ambient
magnetic field has been ignored in this calculation. In cata­
clysmic variables magnetic fields are known to substantially
alter the dynamics of accretion streams (King 1983). Since
in Algol­type binaries the magnetic fields are generally small
compared to cataclysmic binaries it is unlikely they have a
significant effect in directing material associated with mass
transfer. However the formation of primary component hot
spots which may be due to magnetic confinement of accre­
tion streams have been reported for a few Algol systems
(Kappelman & Walter 1979; Ammann & Walter 1973; Wal­
ter 1976) but it is unclear if this mechanism could produce
the 0.23 MeV electron population required to explain the mi­
crowave emission. Florkowski (1980) proposed a mechanism
for Algol whereby material leaks from the L3 Lagrangian
point and forms an impact site with the ambient medium
of ionized hydrogen. The acceleration of particles in this re­
gion can give rise to gyrosynchrotron emission. However this
mechanism requires some correlation with the flaring activ­
ity of the active star and assumes an appreciable amount of
material in a dense accretion disk. This mechanism is again
unlikely for V505 Sgr since such a high density circumstellar
disk has not been observed and because the eclipses of such
an impact site would not be centred at conjunctions.
5.2.4 Unipolar inductor
Significant stressing of the magnetic field of the active star is
possible by analogy with the unipolar inductor mechanism
proposed for the cataclysmic variable AM Her by Chan­
mugam & Dulk (1983) and further investigated by Dulk,
Bastion & Chanmugam (1983). The mechanism proposes
that the primary is embedded in the magnetosphere of the
active star. If the system has some degree of rotational asyn­
chronism then as the field lines cross the surface of the pri­
mary large potential differences are set up across the flux
tube connecting the two stars. This drives an electric cur­
rent from the companion to the polar region of the active
star and back again. If this effect takes place then accelera­
tion of particles in the intra­binary region is possible. Such
a mechanism has not previously been suggested for systems
containing non­degenerate stars because the required mag­
netic fields are many orders of magnitude greater than in­
ferred from the gyrosynchrotron emission from these stars.
Another problem with this process is that it requires some
degree of asynchronism between the two components which
is unlikely for such a closely locked system. Even with re­
duced magnetic fields and high degrees of asynchronism the
required acceleration of electrons to ¸1 MeV is probably not
possible.
c
fl 0000 RAS, MNRAS 000, 000--000

Evidence for an intra­binary radio emission region in the Algol system V505 Sagitarii 9
5.2.5 Field shearing
The remaining mechanism of particle acceleration is perhaps
the most simple because it does not require any ad hoc as­
sumptions on the dynamics of the intra­binary region or the
characteristics of the primary component. This process is
that the non­magnetic companion star alters the magnetic
field topology of the active component in the intra­binary re­
gion. A model of the inductive generation of fields in binary
systems has been presented by Dolginov & Urpin (1978)
and expanded by Dolginov & Urpin (1979) and Dolginov
(1980). Essentially the mechanism gives field excitation in
one star by the currents induced by the second. For the
mechanism to work the rotation axes of the stars must be
slightly inclined to the orbital plane. Although the evolution
of close binaries can result in the alignment of rotation axes
the relative time­scales for synchronization and field induc­
tion are difficult to quantify. The magnetic field topology of
the active component in V505 Sgr is certainly quite com­
plex by solar analogy. However if a dipole model is assumed
and the active star has surface fields Bo then the field near
the primary component is approximately Bo(Rc=d) 3 where
Rc is the radius of the secondary and d is the separation.
The field penetrating into the surface of the primary can
increase due to twisting of the magnetic force lines by an
amount ¸ !R 2
c =j where ! is the angular rotation rate of
the primary and j is the turbulent magnetic viscosity of the
intra­binary plasma. This field in turn induces a field in­
crease near the secondary of order Bo(Rc=d) 6 (!R 2
c =j). The
primary can also increase the magnetic field strength by a
further
amount\Omega R 2
c =j
(where\Omega is the orbital angular veloc­
ity) due to orbital twisting of the field lines. This inductive
generation effect should work very well for close binaries par­
ticularly if they are sufficiently evolved that the induction
dynamo has reached its maximum. This model is strictly
most efficient for a close binary with a dipole field and with
rotation axes not perfectly aligned with the orbital plane.
Dolginov & Urpin (1979) suggest it is most applicable to U­
Geminorum binaries, symbiotic systems and possibly dwarf
binaries.
Both the induction models described above assume a
dipolar field in the magnetic star and also that some mech­
anism exists for inducing currents between the two stars.
In the unipolar inductor model of Chanmugam (1983) the
two stars have parallel rotation axes but are asynchronously
rotating. In the field induction model of Dolginov & Urpin
(1979) both rotational asynchronism and non­parallel ro­
tation axes are considered. Unfortunately for close binary
systems such as Algols neither of these departures from syn­
chronized behaviour are likely. However since the field topol­
ogy in these systems is much more complex and dynamic
it can be proposed, by analogy, that substantial magnetic
amplification can occur in the intra­binary region without
invoking a magnetic field for the A2V star. Rather than ro­
tational effects the predominant factors may be plasma mo­
tions in the lower atmospheres of both stars and magnetic
reconfiguration associated with energy release in the active
regions. The photosphere of the inactive star could play a
crucial role by cutting magnetic field lines rooted there, thus
forming reconnection in the intra­binary region. The subse­
quent particle deposition events could be the source of the
radio emission emanating from between the two stars.
6 CONCLUSIONS
A strong detection of V505 Sgr has been made with fluxes of
¸1.53 mJy and ¸1.37 mJy at 6­cm and 3.6­cm respectively.
The radio flux variations showed the existence of a decrease
of approximately 75% at both conjunctions of the system
and these are best modelled as a large corona surrounding
the active secondary component and a substantial region
of emission originating between the two stars. The possi­
bility that the radiation originates from thermal plasma in
an accretion stream between the stars is discounted. A gy­
rosynchrotron emission process is considered the most likely
explanation with a brightness temperature of Tb ¸ 2.7 10 9
K, a magnetic field strength of 17 Ÿ B Ÿ 170 and an elec­
tron density of ne ¸ 3.7 10 8 cm \Gamma3 . These results are in good
agreement with observations of this system by Umana et al.
(1993).
Umana, Catalano & Rodono (1991) and Umana et al.
(1993) also interpreted the radio spectra of a small sample
of Algols at three frequencies using a core­halo model. They
concluded the emission process was non­thermal in origin
and showed that the core components were smaller than the
size of the active stars with average magnetic fields of 60­
100 G. The halo components were found to be comparable
to the binary system with an average magnetic field of 10­
20 G. These results show that Algols have many similarities
with the observed microwave properties of RS CVn bina­
ries. The present observations are broadly consistent with
this picture although it should be realized that the visibil­
ity data for V505 Sgr show similar average amplitudes on
all baselines for all orientations of the array and so do not
require a core­halo morphology. The existence of eclipsing
behaviour associates the emission with the intra­binary re­
gion and it is possible that this can be further associated
with the hitherto unknown location of the core components
in Algols. These observations cast doubt on the assertion
that the interacting field mechanism for RS CVn binaries
proposed by Uchida & Sakurai (1983) can be discounted
merely on the basis that RS CVns and Algols appear to
show very similar properties. These results not only demon­
strate that field interaction may be important but that it
appears to be important in the very systems in which they
should not exist. This then requires a radical re­assessment
of the details of radio emission processes in active close bi­
nary stars. It is possible that the core structures seen in
Algols and RS CVns can be associated with activity in the
intra­binary region for both classes of object. However this
assertion would require a great deal of additional observa­
tional and theoretical work before being confirmed.
It has been pointed out that binary stars are more likely
to be radio sources than single stars with similar properties
(Gibson 1980). The results of this work may provide some
clues as to why this should be. A correct picture of the radio
activity in close binary stars may then require a detailed as­
sessment of tidal interactions , magnetic field interactions (or
induction) and the effects of circumstellar material. In terms
of the solar paradigm the radio coronae of active close bi­
nary stars seems to be analogous to loop models of emission
but the energetics and geometry may be radically different.
In order to fully understand the origin of radio coro­
nae in active stars further observational work is required.
Multi­wavelength observations, including circular polariza­
c
fl 0000 RAS, MNRAS 000, 000--000

10 A. G. Gunn et al.
tion measurements are the only way forward in this field
since these will enable the appropriate radio emission mech­
anism to be established and provide important constraints
on the physical properties of the radiating plasma. Although
large scale radio coronae are irrefutably established in ac­
tive close binary stars further time­resolved observations of
eclipsing systems are needed to investigate the extent and
location of these regions. In particular it is vital that the
existence of intra­binary emission regions in Algol­type sys­
tems, and perhaps also in RS CVns, is confirmed with ad­
ditional observations. A great deal of work is required in
order to explain the interaction and/or induction of mag­
netic fields in late­type binaries which may be the origin of
some of the observed structures. Overall this field presents
some interesting challenges for the future.
Acknowledgements
Research at Armagh Observatory is grant aided by the Dept.
of Education for N. Ireland. We also acknowledge the com­
puter support by the STARLINK project funded by the UK
PPARC. AGG would like to thank Armagh Observatory for
a research scholarship during the period of this research.
We would also like to thank the ATNF director and staff
(both in Epping and Narrabri). This work was supported
by PPARC Grant No. GR/K60152.
REFERENCES
Abbott D. C., 1985, in Hjellming R. M., Gibson D. M., eds, Radio
Stars. Reidel: Dordrecht, p. 61
Abt H. A., Bidelman W. P., 1969, ApJ, 158, 1091
Ammann M., Walter K., 1973, A & A, 24, 131
Arevalo M., Antonopoulou E., Lazaro C., 1993, in Proceedings of
the First Panhellenic Astronomical Meeting. Athens, p. 267
Batten A. H., 1981, in Kahn T. D., ed, Investigating the Universe.
Reidel: Dordrecht, p. 207
Blondin J. M., Richards M. T., Malinowski M. L., 1995, ApJ, 445,
939
Borra E. F., Landstreet J. D., Mestel L., 1982, ARA & A, 20, 191
Budding E., 1986, Ap & SS, 118, 241
Chambliss C. R., Walker R. L., Karle J. H., Snodgrass H. B.,
Vracko Y. A., 1993, AJ, 106, 2058
Chambliss C. R., 1972, AJ, 77, 672
Chambliss C. R., 1992, PASP, 104, 663
Chanmugam G., Dulk G. A., 1983, in Livio M., Shaviv G., eds,
Cataclysmic Variables and Related Objects, IAU Coll. 72.
Reidel: Dordrecht, p. 223
Clark B. G., Kellerman K. I., Shaffer D., 1975, ApJ, 198, L123
Dempsey R. C., Linsky J. L., Fleming T. A., Schmitt J. H. M.
M., 1993, ApJS, 86, 599
Doiron D. J., Mutel R. L., 1984, AJ, 89, 430
Dolginov A. Z., Urpin V. A., 1978, Sov. Astron., 22, 44
Dolginov A. Z., Urpin V. A., 1979, A & A, 79, 60
Dolginov A. Z., 1980, Sov. Astron. Lett., 6, 132
Doyle J. G., van den Oord G. H. J., Byrne P. B., 1989, A & A,
193, 229
Drake S. A., Abbott D. C., Bastian T. S., Bieging J. H., Church­
well E., Dulk G., Linsky J. L., 1987, ApJ, 322, 902
Drake S. A., Simon T., Linsky J. L., 1989, ApJS, 71, 905
Drake S. A., Simon T., Linsky J. L., 1992, ApJS, 82, 311
Dulk G. A., Marsh K. A., 1982, ApJ, 259, 350
Dulk G. A., Bastion T. S., Chanmugam G., 1983, ApJ, 273, 249
Dulk G. A., 1985, ARA & A, 23, 169
Elias N. M., Gudel M., 1993, AJ, 106, 337
Elias N. M., Mutel R. L., 1993, in Sahade J., McCluskey G. E.,
Kondo Y., eds, The Realm of Iteracting Binary Stars. Kluwer
Academic Publishers, p. 335
Estalella R., Paredes J. M., Rius A., Marti J., Peracaula M., 1993,
A & A, 268, 178
Faraggiana R., Gerbaldi M., Bohm C., 1990, A & A, 235, 311
Florkowski D. R., 1980, in Plavec M. J., Popper D. M., Ulrich R.
K., eds, Close Binary Stars: Observations and Interpretation,
IAU Symp. 88. Reidel: Dordrecht, p. 229
Gary D. E., Linsky J. L., Dulk G. A., 1982, ApJ, 263, L79
Gibson D. M., 1980, in Plavec M. J., Popper D. M., Ulrich R.
K., eds, Close Binary Stars: Observations and Interpretation,
IAU Symp. 88. Reidel: Dordrecht, p. 31
Gimenez A. et al., 1990, in Rolfe E. J., ed, Evolution in Astro­
physics: IUE Astronomy in the Era of New Space Missions.
ESA SP­310, p. 383
Giurkin F., Mardirossian F., Mezzetti M., 1983, ApJS, 52, 35
Gunn A. G., Migenes V., Doyle J. G., Spencer R. E., Math­
ioudakis M., 1997, MNRAS, 287, 199
Hall D. S., 1976, in Multiple Periodic Variable Stars, W. S. Fitch
(ed), IAU Coll. 29, Reidel: Dordrecht, p.287
Hauck B., 1986, A & A, 154, 349
Hoffmeister C., 1934, Astr. Nachr., 251, 321
Horak T., 1967, BAIC, 18, 331
Horak T., 1968, BAIC, 19, 149
Kappelmann N., Walter K., 1979, A & ASS, 38, 161
Khalesseh B., Hill G., 1991, A & A, 244, 75
King A. R., 1983, in Livio M., Shaviv G., eds, Cataclysmic Vari­
ables and Related Objects, IAU Coll. 72. Reidel: Dordrecht,
p. 181
Kitamura M., 1980, in Plavec M. J., Popper D. M., Ulrich R.
K., eds, Close Binary Stars: Observations and Interpretation,
IAU Symp. 88. Reidel: Dordrecht, p. 95
Kwee K. K., 1953, BAN, 12, 35
Landstreet J. D., 1982, ApJ, 258, 639
Lause F., 1935, Astr. Nachr., 258, 311
Lause F., 1938, Astr. Nachr., 266, 17
Lestrade J. F., Mutel R. L., Preston R. A., Phillips R. B., 1988,
ApJ, 328, 232
Lestrade J. F., 1988, in Moran J., Reid M., eds, The Impact of
VLBI on Astrophysics and Geophysics. Reidel: Dordrecht, p.
265
Lestrade J. F., 1996, in K. G. Strassmeier & J. L. Linsky (eds),
Stellar Surface Structure, IAU Symp. 176. Kluwer Academic
Publishers, p. 173
Linsky J. L., 1990, in Ibanoglu C., ed, Active Close Binaries.
Kluwer: Dordrecht, p. 747
Lubow S. H., Shu F. H., 1975, ApJ, 198, 383
MacAlister H. A., Hartkopf W. I., Hutter D. J., Shara M. M.,
Franz O. G., 1987, AJ, 93, 183
Magalashvili N. L., Razmadze N. A., 1953, Peremennye Zvezdy,
10, 313
McCluskey G. E., 1993, in Sahade J., McCluskey G. E., Kondo Y.,
eds, The Realm of Iteracting Binary Stars. Kluwer Academic
Publishers, p. 39
Melrose D. B., Dulk G. A., 1982, ApJ, 259, 844
Mezzetti M., Cester B., Giurcin G., Mardirossian F., 1980, A &
ASS, 39, 273
Morgan W. W., Keenan P. C., Kellman E., 1943, An Atlas of
Stellar Spectra. University of Chicago Press, Chicago
Morris D. H., Mutel R. L., 1988, AJ, 95, 204
Mutel R. L., Doiron D. J., Lestrade J. F., Phillips R. B., 1984,
ApJ, 278, 220
Mutel R. L., Lestrade J. F., Preston R. A., Phillips R. B., 1985,
ApJ, 289, 262
Olson E. C., 1985, in Eggleton P. P., Pringle J. E., eds, Interacting
Binary Stars. Reidel: Dordrecht, p. 127
c
fl 0000 RAS, MNRAS 000, 000--000

Evidence for an intra­binary radio emission region in the Algol system V505 Sagitarii 11
Owen F. N., Spangler S. R., 1977, ApJ, 217, L41
Popper D. M., 1949, ApJ, 109, 100
Pringle J. E., 1985, in Pringle J. E., Wade R. A., eds, Interacting
Binary Stars. Cambridge University Press, p. 1
Richards M. T., 1990, ApJ, 350, 372
Rovithis­Livaniou H., Rovithis P., 1992, IBVS, No. 3803
Rovithis­Livaniou H., Rovithis P., 1994, A & ASS, 105, 47
Rovithis­Livaniou H., Rovithis P., Antonopoulou E., Kalimeris
A., 1991, IBVS, No. 3608
Russell H. N., Merrill J. E., 1952, Technical Report, Princeton
University
Schnopper H. W., Delvaille J. P., Epstein A., Heimken H., Murray
S. S., Clark G., Jernigan G., Doxsey R., 1976, ApJ, 210, L75
Shu F. H., Lubow S. H., 1981, ARA & A, 19, 277
Simon T., Linsky J. L., Schiffer F. H., 1980, ApJ, 239, 911
Singh K. P., Drake S. A., White N. E., 1996, AJ, 112, 221
Slee O. B., Nelson G. J., Stewart R. T., Wright A. E., Innis J. L.,
Ryan S. G., Vaughan A. E., 1987, MNRAS, 229, 659
Slonim J. M., 1934, Astr. Nachr., 253, 367
Sofronitsky A. V., 1953, Izv. Pulkovo Obs., 19, 1
Stephan C. P., 1978, IBVS, No. 1502
Stern R. A., Haisch B. M., Nagase F., Uchida Y., Tsuneta S.,
1990, in Wallerstein G., ed, Cool Stars, Stellar Systems and
the Sun. ASP Conference Series Vol. 9, p. 224
Stern R. A., Uchida Y., Tsuneta S., Nagase F., 1992, ApJ, 400,
321
Stewart R. T., Slee O. B., White G. L., Budding E., Coates D.
W., Thompson K., Bunton J. D., 1989, ApJ, 342, 463
Swank J. H., Holt S. S., White N. E., Becker R. H., 1981, ApJ,
246, 208
Tomkin J., 1992, ApJ, 387, 631
Turner K. C., 1985, in Hjellming R. M., Gibson D. M., eds, Radio
Stars. Reidel: Dordrecht, p. 283
Uchida Y., Sakurai T., 1983, in Rodono M., Byrne P. B., eds,
Activity in Red Dwarf Stars, IAU Coll. 71. Reidel: Dordrecht,
p. 629
Umana G., Catalano S., Rodono M., 1991, A & A, 249, 217
Umana G., Trigilio C., Hjellming R. M., Catalano S., Rodono M.,
1993, A & A, 267, 126
Vracko Y. A., Snodgrass H. B., Karle J. H., 1990, BAAS, 22, 831
Wade M., Hjellming R. M., 1972, Nature, 235, 270
Walker R. L., 1993, AJ, 106, 2051
Walter F. M., Cash W., Charles P. A., Bowyer C. S., 1980, ApJ,
236, 212
Walter K., 1976, A & ASS, 26, 227
Walter K., 1981a, A & A, 101, 369
Walter K., 1981b, A & ASS, 46, 263
White N. E., Marshall F. E., 1983, ApJ, 268, L117
White N. E., Culhane J. L., Parmar A. N., Kellett B. J., Kahn
S., van den Oord G. H. J., Kuijpers J., 1986, ApJ, 301, 262
Wolf G. W., Kern J. T., 1983, ApJS, 52, 429
Wolf G. W., Kern J. T., Hayes T. L., Chaffin C. R., 1982, IBVS,
No. 2185
Wood D. B., 1972, Technical Report, NASA Publ. No. X­110­72­
473
Zirin H. Astrophysics of the Sun, p. 134, Cambridge University
Press, 1988
c
fl 0000 RAS, MNRAS 000, 000--000