THE 1998 MEETING OF THE VARIABLE STAR SECTION TRISTRAM BRELSTAFF The 1998 meeting of the BAA-VSS was held on Saturday 31st October at Crayford Manor House at Crayford in Kent. The meeting was opened by Roger Pickard who welcomed the participants and then introduced the first speaker, Dr Robert Smith of Sussex University. Dr Smith explained that he had spoken on the subject of 'Professional-Amateur Symbiosis' at the BAA-VSS Centenary Meeting in 1991 and that he was going to continue on this theme by giving a brief history of two of the observatories at Sussex University. The first was built after a Mr Madge of Burgess Hill donated a 6-inch refractor to the university in 1964. However, in 1973 some students 'swapped' this telescope for a 10-inch reflector from a well-known telescope supplier (Dr Smith stressed that this was done without his knowledge!). During the 1980's light-pollution and vandalism at the site became increasingly worse and in 1998 the dome was abandoned. The second observatory was built after the university bought an 18-inch reflector from a Mr H. Robin of Tunbridge Wells in 1990. Difficulties with planning permission meant that the observatory was not opened until 1996. It is equipped with a CCD but, because of lack of man-power, it is not being used for 'real' research, only used for undergraduate projects. However, it has been used to measure the orbit of T Tauri. Next, the BAA-VSS Director, Gary Poyner, reviewed what had been happening in the BAA-VSS over the past 12 months. For 1997 more than 52000 observations of 541 stars had been submitted by about 700 observers (these figures include observations received from the SPA-VSS and the RASNZ-VSS, as well as observations submitted to The Astronomer magazine). Gary showed several slides to show how the observations were distributed between the various programs. As usual, the bulk of the observations were of stars on the main 'Telescopic' and 'Binocular' programs. Most of the 'Telescopic' program observations were of cataclysmic variables - over 16000 were of dwarf novae alone. Three quarters of the 'Binocular' program observations were of semiregular variables. 6500 observations of stars on the 'Recurrent Objects Program' were received. Gary was a bit disappointed that more CCD observers weren't monitoring these stars. Neither the Active Galaxy Program (run in conjunction with the Deep-Sky Section) nor the Mike Collins' Star Program were very popular during the year. Gary went on to describe some special projects that BAA-VSS members had been involved in during 1997. Six dwarf novae were monitored for Dr W. Skidmore of Keele University who was hoping to use the William Herschel Telescope to look for oscillations at minimum. Unfortunately, all six stars were in outburst during his observing run. EM Cygni was monitored for outbursts and standstills at the request of Les Thomas (Keele). Eclipses were observed in outburst. Observations of 3C66A and OJ287 were contributed to the 'Blazar 2000 Project' coordinated by Harry Lehto and Dr Mark Kidger. Observations were also contributed to the IP Pegasi campaign coordinated by Dr Tim Naylor (also Keele). Gary mentioned Bill Worraker as having done a lot of work on this. In rounding off, Gary mentioned that John Greaves had been comparing some of our comparison star magnitudes against the Hipparcos and Tycho Catalogues. This sparked off a discussion. Norman Walker said that he had been comparing Kevin West's photoelectric results against Tycho magnitudes and had found what he thought was a periodic error of about 0.2 mag in the Tycho magnitudes. However, he explained that it was fairly easy to get an internal accuracy of 0.01 mag within a particular comparison star sequence, and this was all that we needed for our purposes - we do not need to tie the magnitudes into a consistent all-sky system. Dr Richard Miles said that below mag 8 - 9 the Tycho data shows errors of 0.1 - 0.2 mags. At this point Roger Pickard brought the discussion to an end by suggesting that we should consider organising a separate meeting just for this subject! Then Eric Broens of the VVS (the Belgian variable star organisation) rose to speak on the period history of Chi Cygni. It was a few years ago that he had first constructed an O-C diagram for this star using VVS and Dutch observations covering the years 1959 - 1994. He then started working with Chris Sterken and Chris Koen and had obtained observations from the BAA-VSS, from Japan and from other sources which extended the coverage back to 1838. They determined times of maximum by fitting 3rd degree polynomials and then used linear regression to determine the best-fit period. The resulting diagram showed large, possibly cyclic variations in O-C ranging from +40d to -40d. Eric explained that such apparently cyclic variations in O-C do not necessarily indicate the presence of true period variations. Back in the 1920's Eddington and Plakidis had pointed out that random variations in period length in Mira stars could easily accumulate to give the appearance of systematic variations in the period. In the 1930's, Sterne showed that it was necessary to use statistical techniques to assess the reality of such period changes. Analysis of the Chi Cygni data by Koen (using a technique developed by Koen and Lombard) indicated that the observed variations were only marginally significant (at the 90% level). However, Chris Sterken then managed to obtain copies of observations by G. Kirch, the discoverer of Chi Cygni, C. Kirch, Cassini and Halley covering the years 1686 - 1736. These observations roughly doubled the timespan of the O-C diagram and revealed the presence of a systematic increase in period of 0.013d per year (equivalent to 4d since 1868). This period change is significant (at the 95% level). There are even suggestions in the earlier data of the cyclic variations suspected in the post-1838 data. The increase in the period could be the result of the star having undergone a 'helium shell flash'. This is thought to occur in red giant stars with an inert carbon-oxygen core surrounded by an inert helium shell and, outside that, a burning hydrogen shell. As the core slowly contracts, the helium shell ignites, expands and then stops burning again at intervals of tens of thousands of years. Modelling by Wood and Zarro (1981) indicates that these flashes are accompanied by changes in luminosity and in pulsation period. Eric showed graphs of how these would be expected to vary from one flash to the next, and indicated the positions of R Aql, R Hya, W Dra and T UMi, some Mira stars known to have strongly varying periods. T UMi, in particular, is thought to have undergone its flash only very recently. Chi Cygni, on the other hand, probably underwent its about 10000 years ago and is now in the recovery phase, with most of its energy coming from burning in the hydrogen shell. Eric said that a paper these results had just been accepted for publication in Astronomy & Astrophysics (Sterken, Broens & Koen, 'On the Period History of Chi Cygni') and that the data would appear in the Journal of Astronomical Data (a CD-ROM journal). In the discussion, Albert Zijlstra suggested that it might be possible to explain the observed period change by mass-loss alone, without a helium-flash. The last speaker of the morning was Guy Hurst who described his experiences in CCD Nova Patrolling. He bought a Starlight Xpress MX5 CCD from Terry Platt and fitted a standard camera lens to it (he was quoted £80 for a spacer tube to do this but managed to find one for £10 in a second-hand shop). He first practised taking images of a street-lamp before turning to the night sky. He showed an image of the Pleiades he had obtained. Initially he had had no finder so he had to point it in the direction of a bright star and hope for the best. He showed an image of the region of Gamma Cygni. He found that with a 5 second exposure the image trailing was not too bad and that by combining 4 such images he could reach mag 12. In order to search for novae he first produces a set of master images of the area he is to search and then uses these to compare later images against. He decided not to use atlases for reference purposes as it is always better to use reference images taken with the same instrument that is to be used in the search. Guy feels that his search method must be effective because it is finding lots of previously-known variables in the mag 8 - 10 range. In the discussion Guy said that the images were store on the hard disc of a lap-top computer. Dr Richard Miles suggested that Guy should drive his CCD piggy-back on a Meade but Guy said that one of the ideas behind his project was that it should be low-cost! After lunch, as the first scheduled speaker, Tonny Vanmunster had not yet arrived, Paul Roche of Sussex University spoke on the subject of 'Echo-sounding the Accretion Disk of Her X-1'. Paul is a member of an international team attempting to determine the mass of the neutron star in the Her X-1 system. The masses of neutron stars are currently a subject of great theoretical interest. On the one hand, we have very well-determined masses for those neutron stars in binary pulsars, all of which are close to 1.45 solar masses. On the other hand, we only have rather inaccurate masses for the neutron stars in X-ray binaries, and these show a much greater range with some, such as Her X-1, possibly being as small as 0.7 solar masses. The aim of the international team is to use high time-resolution observations to model the Her X-1 system, and to derive a more accurate mass for the neutron star. Her X-1 is the X-ray counterpart of theoptical variable star HZ Her. The Her X-1/HZ Her system is thought to be a binary consisting of a relatively 'normal' A-type star ('HZ Her'), which produces most of the optical radiation of the system, the neutron star ('Her X-1') which produces the X-rays, and an accretion disk surrounding the neutron star which is visible in the ultra-violet. Paul explained that some of the key inputs to the model were parameters defining the structure and orientation of the disk. The hoped to determine these parameters by observing ultra-violet 'echoes' of the X-ray pulses from the neutron star. These X-ray pulses are emitted at intervals of 1.24s, and are relatively narrow. However, some of the energy from them is absorbed by the disk and re-radiated as 'echoes' in the ultra-violet and, as each pulse takes longer to reach the outer parts of the disk than the inner parts, the echoes we see are broader than the original pulses. It is this broadening that they are using to constrain the structure and orientation of the disk in their models. To do this they need lots of high time-resolution observations at many wavelengths. In 1998 they were allocated large amounts of observing time on the Rossi X-ray Timing Explorer, and Extreme Utra-violet Explorer satellites, the Hubble Space Telescope, as well as on the Keck, William Herschel, Calar Alto and Mount Palomar Telescopes. However, further data is required and more observations are planned for 1999. Simultaneous optical monitoring of HZ Her by amateurs would be very useful. Paul added that preliminary modelling suggested that the mass of the neutron star was in the region of 1.5 solar masses, in agreement with the binary pulsar results. The next speaker was Zach Ioannou of Keele University who spoke on 'Outbursts of Dwarf novae: Science with Medium-size Telescopes'. Zach had been modelling the dwarf novae in outburst using CCD observations made by various professional astronomers at Keele and St Andrews, as well as by the amateurs David Strange and Nick James. The first star he discussed was HT Cas which had been observed during its 1995 outburst. The eclipses are narrow at minimum, become broader at maximum and then narrow again on the fade. The development of the eclipses during the outburst is best modelled by an 'outside-in' type of outburst, in which the rim of the disk brightens first and the brightening then moves inwards. Timings of eclipses on three consecutive nights also show a peculiar 80s deviation in O-C on the second night, which is rather difficult to explain by any of the available models. The second star Zach discussed was IP Peg, which was observed during its 1997 outburst. This star shows markedly asymmetric eclipses at minimum: the fade is very steep and then the rise starts slowly but then suddenly becomes much steeper. In addition, the out-of-eclipse level before the eclipse is noticably brighter than the one after. Zach interpreted these as features due eclipses of both the white dwarf and the hot-spot. As the outburst starts, the eclipse becomes more symmetrical, and the rise straightens out somewhat. There is also a suggestion of a secondary eclipse at phase 0.5 in some of the observations. By the time the outburst has reached maximum the eclipses have become broad and shallow and there is now a marked hump at phase 0.5. Outside of eclipse there are flares, which suggests t he presence of tidal instability in the disk. Modelling suggests that the disk is stongly flared (ie: it has a high outer rim) and that the outbursts are of the 'outside-in' variety. The disk seems to undergo quite dramatic changes during the outbursts of this star. Then Andy Hollis, the director of the BAA Asteroids and Remote Planets Section, rose to talk about 'Automating a Telescope for Photometry'. Andy had recently finished converting his 14 and 17-inch Cassegrain telescopes, which he had previously been using for manual photoelectric photometry, over to automatic operation. Both of these telescopes are mounted on the same German mount. They are controlled with a BBC Micro running software written by Peter and Ann Ells for the Jack Ells APT. However, Andy has adapted the software to work with DC rather than pulse counting techniques, and has also made a few other changes to streamline its operation. Since completing the conversion, he had only had one night during which he could observe. This, he had used to monitor two constant stars near Procyon. The resulting measurements had an accuracy of about 0.02 mags with only one or two wild points (these Andy blamed on aircraft condensation trails). Andy stressed the importance of having a systematic approach to testing and correcting the system. In particular, he warned against making adjustments in the middle of observing runs. Instead you should keep a log of any problems noticed and only make the required adjustments between runs. Andy summed up by saying that the great advantage of his new automated set-up was that it only took him 15 minutes to set it up at the start of a run and another 15 minutes to close it down at the end - in between he was free to do other things, like sleep!. The last speaker before the tea break was Nick James who spoke in the subject of 'Linux and the Variable Star Observer'. Nick had recently been appointed the 'official' distributor of the Starlink CD-ROM for the amateur community. The Starlink software collection is a suite of astronomical software that is used by UK professional astronomers to perform image analysis, time series analysis and various other tasks. The original Starlink software ran under Unix (a computer operating system popular with academics) but the software has been ported to Linux (a version of Unix suitable for running on PCs), and it is this version that is distributed by Nick. Nick showed some screen-shots of PISA, one of the Starlink programs. 'PISA' stands for 'Position, Intensity, Shape Analysis' and this program can be used to identify all the star images in a CCD image and produce a file of their intensities and x-y coordinates. Nick said that on his 200MHz Pentium PC it could process an image of M13 in about 2s. Because the Starlink programs have been produced and extensively used by professional astronomers, they tend to be based on the best-available algorithms and be well-adapted to handle large data sets. In addition, they usually give error estimates on the results they produce. They are also generally well-documented. If you would like to run Starlink programs on your own PC then you will ned to install 'Red Hat' Linux. This will probably mean partitioning your hard disk and then installing Windows in the lower partition and Linux in the upper one (Nick pointed out that when you buy a new PC your supplier should be able to do this for you). You will then be able to dual-boot into Window or Linux as you require. Another option is to buy a second PC just for Linux. One of the great advantages of Linux is that it is free. Another is that it is relatively small, however it is still too large to download over the Internet. After the tea break, Tonny Vanmunster gave his talk on "EG Cancri and the Center for Backyard Astrophysics (CBA)". The CBA is a group of professional and amateur astronomers who use CCDs to study the short-period photometric cycles of cataclysmic variables. Tonny first described his own observatory, which he referred to as 'CBA-Belgium'. He had two Schmidt-Cassegrain telescopes (35cm and 25cm) and computer hardware for recording the CCD images. At the start of each clear night he sets up the larger telescope to automatically track a particular star of interest and record images every minute or so. At the end of the night he uses software to derive magnitudes from the recorded images. Since 1996 he had done 1279 hours of photometry in this way and had obtained some 48000 usable images in the process. Tonny then described EG Cancri. This star is normally about mag 19 but had been found in outburst at mag 13 in 1977 by Huruhata. There was also an unconfirmed report of it at mag 13.4 in 1995 by Watanabe. And that was all, that is until 1996 November 30th when Patrick Schmeer found it again in outburst at mag 12. This discovery sparked a campaign of intensive observations by the CBA and other groups. The resulting light-curve showed a characteristic SU UMa type superoutburst, lasting about 20 days, followed closely by a series of six 'echo' outbursts at intervals of about 6 days. These 'echo' outbursts were smaller and narrower than the main one and reached only mag 13-14. Photometry by the CBA team showed the presence of 'superhumps' with a period of about 0.06d at the maximum of the main outburst. This is unusually early for superhumps - normally they don't develop until about 5 days into an outburst. Their amplitude then increased as the outburst developed but they faded away again on the fade. Superhumps were not present on the first few echo outbursts but they did appear again on the later ones. Spectroscopy during the outburst confirmed the star to be a dwarf nova. The spectrum at minimum was very similar to the minimum spectrum of WZ Sge, another dwarf nova which only erupts every few decades. An early onset of superhumps was also observed in WZ Sge and in the WZ Sge-like stars AL Com and HV Vir. Maybe this is characteristic of these stars? However, the strangest feature of the 1996-97 outburst of EG Cnc was undoubtedly the six 'echo' outbursts. Echo outbursts have been seen before - in UZ Boo, T Leo and GO Com - but only one or two at a time. Tony mentioned two theories which have been proposed to explain similar echo outbursts seen in X-ray sources. In one theory they are due to episodic accretion from the disk onto the compact component. Each such accretion event heats up the primary enough to cause it to deposit more matter into the disk . This matter then takes about 6 days to spiral in to the inner edge of the disk where it causes the next accretion event. In the other theory the echo outbursts are due to heating waves 'reflected' from cooling waves accompanying the decline from the previous outburst. Next Albert Zijlstra of UMIST (University of Manchester Institute of Science and Technology) spoke on "The Hesitant Mira R Doradus". He started off by discussing planetary nebulae. First he showed an image of the Egg Nebula in which the central star is invisible because it is undergoing catastrophic mass loss. Such a star is probable suffering from something like 30 mags of extinction in visible light - rather like forgetting to open the dome of your obserevatory! The Egg Nebula could soon become a planetary nebula. A few years ago Shklovsky had suggested that if you were to compress a planetary nebula back onto its central star then what you would be left with would be a Mira star. Albert showed an image of NGC 7027, saying that this could well have been a Mira star just a thousand years ago. The catastrophic mass loss, or super-wind, that leads to the formation of planetary nebulae is thought to be related to the mass loss from red giant stars, especially Mira stars. Mira stars are pulsating red giant stars with amplitudes greater than 2.5 mags and periods of about 200 - 500d. This makes them easy for amateurs to observe but awkward for professionals (they would prefer 2 - 5d!). Semiregular (SR) variables are similar but have amplitudes less than 2.5 mags. There are also the so-called OH/IR stars which have periods of 700 - 3000d and are only 'visible' in the infra-red. An evolutionary scheme has been proposed that goes: SR -> Mira -> OH/IR -> PN, however, Albert added that this 'was so neat it cannot be true'. Turning to R Dor, this star has a range of mag 5 - 7 and a period of 300d. It is very bright in the infra-red, indeed in the K-band it is the brightest star in the sky after the Sun. Its angular diameter at 55 milliarcsecs (mas) is even larger than that of Betelgeuze (40 mas). At the Hipparcos distance of 62pc this corresponds to a radius of 380 solar radii. From these properties one would expect the star to have a period of about 150d. Albert said that he became interested in R Dor when he saw some observations of it by the New Zealand amateur Albert Jones. Jones' observations covered the years 1944 - 1967 and seemed to show a sudden change from Mira-like to SR behaviour in 1952. Before then it had an amplitude of 2 mags, then the period split and this drew power from the main pulsation and the amplitude decreased to that typical of an SR variable. Albert (Zilstra) then searched for other observations to fill the gap since 1967. Unfortunately the AAVSO observations were only available as 10-day means and showed large scatter, so he couldn't use them. However, observations by the RASNZ were available and observer corrections could be applied to to make these acceptable for filling in the gap since 1967. Period analysis shows that there are two periods present in the whole dataset, 330d and 170d, the former is only present in the earlier observations, the latter only in the later ones. This proves that the 170d period is not an alias or harmonic of the 330d one. Albert then showed the results of wavelet analysis. This consisted of a contour plot with time in years along the x-axis, period in days along the y-axis, the contour lines indicating the 'goodness of fit' of each period at each time. This plot showed that the best fit period had flipped discontinuously from 330d to 170d and back again several times since 1944. When the 330d period is present, R Dor is a Mira/SRa variable, when the 170d period is present then it is an SRb variable. Mira stars are thought to pulsate in the fundamental mode (that is all parts of the star move in and out together). However, the above periods along with the Hipparcos distance suggest that R Dor pulsates in the first overtone (the outer layer move out as the inner layers fall in, and vice-versa). For R Dor to be a fundamental mode pulsator it would have to have a mass of about 4 solar masses, which is unacceptably large. Mira stars in the Large Magelanic Cloud with periods less than 400d show a period-luminosity (P-L) relationship rather like that for cepheids. In the past it has been thought that this relationship does not hold in our Galaxy. However, R Dor fits the Mira P-L relationship rather well so this might have to be revised. Albert said that the Mira PL relationship has the potential for being more useful that the cepheid one. The distance to the LMC is 52kpc according to cepheids, but only 45kpc according to RR Lyrae variables and SN1987a. One of the aims of Hipparcos was to get accurate distances to cepheids and RR stars, but in this it failed because the nearest stars of these classes turned out to be just too far away. One way to reconcile the various distances to the LMC is suggested by Mira stars in globular clusters. These stars appear to be anomalous when compared with nearby Mira stars. However, this anomaly would disappear if the cepheid distance scale was slightly wrong and the LMC was 10% further away than current estimates put it. Albert also pointed out that this would also reconcile the the problem of the universe appearing to be younger than the oldest stars. However, this is rather a long way to go based on observations of just one star (R Dor) by just one observer (Albert Jones). To sum up: R Dor suggests that the Mira stars and semiregular variables are closely related and are two sides of the same type. The mass loss history of R Dor suggests that the superwinds are intermittent, not continuous, and are related to the halos around planetary nebulae. The pulsational characteristics of R Dor suggest it is pulsating in the first overtone rather than the fundamental mode. And finally, the Mira period-luminosity relationship could be used to calibrate the cepheid and RR Lyrae distance scales. In answer to a question, Albert said that V Boo was 'similar but not similar to' R Dor, but he recommended continued observation of it. Norman Walker asked about non-radial pulsation modes (as are found in some blue stars). Albert said in red giants these would only result from tidal effects - such as swallowing Jupiter-size planets. The final talk of the day was given by John Howarth of CMHAS (Crayford Manor House Astronomical Society) and was entitled "Analysis of Quasi-periodic Data". John had been working with John Greaves on a study of changes in period, phase and amplitude in Mira and semiregular variables. For this they used observations by the BAA-VSS and other societies. John showed a light-curve of W Cyg which is known to have periods of 130d and 234d. He pointed out how in different parts of the light-curve both periods are either present together or else one is absent and the other present by itself. There is even a stretch of what appears to be chaotic behaviour in one part of the light-curve. John then described the use of the discrete Fourier transform to derive period, phase and amplitude information, and to distinguish true multiple periods from aliases. He illustrated this with some results for T And. He went on to show how a 'moving window' could be used to generate diagrams of the variations in phase and amplitude in T Cas, Chi Cyg and other stars. The phase-time time diagrams correlate closely with the O-C diagrams for these stars and actually show less scatter (they use information from the whole of each cycle - not just the maximum and minimum). For WY Cyg there also seems to be a correlation between the phase and the amplitude. John finished off by showing some phase-ampltude diagrams plotted in polar coordinates. In these the points start to rotate about the origin as the period changes, getting faster and faster as the period gets further and further out of step with the original one.