We present a new edition of the catalogue of high-mass X-ray binaries in the Galaxy. The catalogue contains source name(s), coordinates, finding chart, X-ray luminosity, system parameters, and stellar parameters of the components and other characteristic properties of 114 high-mass X-ray binaries, together with a comprehensive selection of the relevant literature. The aim of this catalogue is to provide the reader with some basic information on the X-ray sources and their counterparts in other wavelength ranges ($\gamma$-rays, UV, optical, IR, radio). About 60% of the high-mass X-ray binary candidates are known or suspected Be/X-ray binaries, while 32% are supergiant/X-ray binaries. Some sources, however, are only tentatively identified as high-mass X-ray binaries on the basis of their X-ray properties similar to the known high-mass X-ray binaries. Further identification in other wavelength bands is needed to finally determine the nature of these sources. In cases where there is some doubt about the high-mass nature of the X-ray binary this is mentioned. Literature published before 1 October 2005 has, as far as possible, been taken into account.

We present a new edition of the catalogue of the low-mass X-ray binaries in the Galaxy and the Magellanic Clouds. The catalogue contains source name(s), coordinates, finding chart, X-ray luminosity, system parameters, and stellar parameters of the components and other characteristic properties of 187 low-mass X-ray binaries, together with a comprehensive selection of the relevant literature. The aim of this catalogue is to provide the reader with some basic information on the X-ray sources and their counterparts in other wavelength ranges ($\gamma$-rays, UV, optical, IR, and radio). Some sources, however, are only tentatively identified as low-mass X-ray binaries on the basis of their X-ray properties similar to the known low-mass X-ray binaries. Further identification in other wavelength bands is needed to finally determine the nature of these sources. In cases where there is some doubt about the low-mass nature of the X-ray binary this is mentioned. Literature published before 1 October 2006 has, as far as possible, been taken into account.

On the basis of spectrophotometric observations we reconstruct the accretion disk of the eclipsing novalike cataclysmic variable RW Tri in the wavelength region 3600-7000 A. We find a radial temperature profile that is, on average, consistent with that expected on the basis of the theory of optically thick, steady state accretion disks and infer a mass-accretion rate in RW Tri of \~10e-8 solar masses per year. The line emission is dominated by two areas: one around the hot-spot region and one near the white dwarf. Both emission regions have appreciable vertical extension, and seem to be decoupled from the velocity field in the disk. In our observations RW Tri shows a number of features that are characteristic of the SW Sex sub-class of novalike stars. The appearance of a novalike system as a UX UMa/RW Tri or SW Sex star seems to be mainly governed by the mass-transfer rate from the secondary at the time of observation.

The connection between Gamma-Ray Bursts (GRBs) and their afterglows is currently not well understood. Afterglow models of synchrotron emission generated by external shocks in the GRB fireball model predict emission detectable in the gamma-ray regime ($\gax 25$ keV). In this paper, we present a temporal and spectral analysis of a subset of BATSE GRBs with smooth extended emission tails to search for signatures of the ``early high-energy afterglow'', i.e., afterglow emission that initially begins in the gamma-ray phase and subsequently evolves into X-Ray, uv, optical, and radio emission as the blast wave is decelerated by the ambient medium. From a sample of 40 GRBs we find that the temporal decays are best described with a power-law $\sim t^{\beta}$, rather than an exponential, with a mean index $<\beta > \approx -2$. Spectral analysis shows that $\sim 20%$ of these events are consistent with fast-cooling synchrotron emission for an adiabatic blast wave; three of which are consistent with the blast wave evolution of a jet, with $F_{\nu} \sim t^{-p}$. This behavior suggests that, in some cases, the emission may originate from a narrow jet, possibly consisting of ``nuggets'' whose angular size are less than $1 / \Gamma$, where $\Gamma$ is the bulk Lorentz factor.

We measured the radial-velocity curve of HD77581, the B-supergiant companion of the X-ray pulsar Vela X-1, using 183 high-resolution optical spectra obtained in a nine-month campaign. We derive radial-velocity amplitudes for different lines and wavelength regions, and find all are consistent with each other, as well as with values found in previous analyses. We show that one apparent exception, an anomalously low value derived from ultra-violet spectra obtained with the IUE, was due to an error in the analysis procedures. We re-analyse all IUE spectra, and combine the resulting velocities with the ones derived from the new optical spectra presented here, as well as those derived from optical spectra published earlier. As in previous analyses, the radial velocities show strong deviations from those expected for a pure Keplerian orbit. The deviations likely are related to the pronounced line-profile variations seen in our spectra. It turns out that systematic deviations as a function of orbital phase are present. Our best estimate of the radial-velocity amplitude, Kopt = 21.7 +/- 1.6 km/s, has an uncertainty not much reduced to that found in previous analyses, in which the systematic deviations had not been taken into account. Combining our velocity amplitude with the accurate orbital elements of the X-ray pulsar, we infer M_ns sin^3i = 1.78 +/- 0.15 Msun.

This paper reports the results of the analysis of the second ROSAT PSPC survey of M31 performed in summer 1992. We compare our results with those of the first survey. Within the ~10.7 deg^2 field of view, 396 individual X-ray sources are detected in the second survey data, of which 164 are new detections. When combined with the first survey, this result in a total of 560 X-ray sources in the field of M31. Their (0.1 keV - 2.0 keV) fluxes range from 7 x 10^-15 to 7.6 x 10^-12 erg cm^-2 s^-1, and of these 560 sources, 55 are tentatively identified with foreground stars, 33 with globular clusters, 16 with supernova remnants, and 10 with radio sources and galaxies (including M32). A comparison with the results of the Einstein M31 survey reveals 491 newly detected sources, 11 long term variable sources, and 7 possible transient sources. Comparing the two ROSAT surveys, we come up with 34 long term variable sources and 8 transient candidates. For the M31 sources, the observed luminosities range from 4 x 10^35 to 4 x 10^38 erg s^-1. The total (0.1 keV - 2.0 keV) luminosity of M31 is (3.4+-0.3) x 10^39 erg s^-1, distributed approximately equally between the bulge and disk. Within the bulge region, the luminosity of a possible diffuse component combined with faint sources below the detection threshold is (2.0+-0.5) x 10^38 erg s^-1. An explanation in terms of hot gaseous emission leads to a maximum total gas mass of (1.0+-0.3) x 10^6 M_sun.

We present and discuss the photometric and spectroscopic evolution of the peculiar SN1998bw, associated with GRB980425, through an analysis of optical and near IR data collected at ESO-La Silla. The spectroscopic data, spanning the period from day -9 to day +376 (relative to B maximum), have shown that this SN was unprecedented, although somewhat similar to SN1997ef. Maximum expansion velocities as high as 3x10^4 km/s to some extent mask its resemblance to other Type Ic SNe. At intermediate phases, between photospheric and fully nebular, the expansion velocities (~10^4 km/s) remained exceptionally high compared to those of other recorded core-collapse SNe at a similar phase. The mild linear polarization detected at early epochs suggests the presence of asymmetry in the emitting material. The degree of asymmetry, however, cannot be decoded from these measurements alone. The HeI 1.083 mu and 2.058 mu lines are identified and He is suggested to lie in an outer region of the envelope. The temporal behavior of the fluxes and profiles of emission lines of MgI]4571A, [OI]6300,6364A and a feature ascribed to Fe are traced to stimulate future modeling work.

We report the discovery of a kilohertz quasi-periodic oscillation (kHz QPO) in the low-mass X-ray binary (LMXB) 2S 0918-549. The kHz QPO has a frequency of 1156+-9 Hz, a FWHM of 70+-20 Hz, and a fractional rms amplitude of 18%+-2%. We also detected for the first time a type I X-ray burst in this source. The compact object must therefore be a neutron star. Finally, we were able to classify the source as an atoll source exhibiting the full complement of spectral/timing states of this class (island and banana). From the peak burst flux an upper limit on the persistent luminosity can be derived of 0.5% of the Eddington luminosity, making 2S 0918-549 one of the least luminous LMXBs showing a kHz QPO. We compare the fractional rms amplitudes of the upper kHz QPO across the ensemble of LMXBs. We find a strong anticorrelation with luminosity. In LMXBs with luminosities ~100 times lower than those of Z-sources, the fractional rms amplitude is a factor ~10 larger.

The Burst and Transient Source Experiment (BATSE) on the Compton Gamma Ray Observatory detected gamma-ray bursts (GRBs) with a real-time burst detection (or "trigger") system running onboard the spacecraft. Under some circumstances, however, a GRB may not have activated the onboard burst trigger. For example, the burst may have been too faint to exceed the onboard detection threshold, or it may have occurred while the onboard burst trigger was disabled for technical reasons. This paper describes a catalog of 873 "non-triggered" GRBs that were detected in a search of the archival continuous data from BATSE, recorded between 1991 December 9.0 and 1997 December 17.0. For each burst, the catalog gives an estimated source direction, duration, peak flux, and fluence. Similar data are presented for 50 additional bursts of unknown origin that were detected in the 25--50 keV range; these events may represent the low-energy "tail" of the GRB spectral distribution. This catalog increases the number of GRBs detected with BATSE by 48% during the time period covered by the search.

We present low-resolution spectrophotometric optical observations of the eclipsing nova-like cataclysmic variable SW Sex, the prototype of the SW Sex stars. We observed the system when it was in an unusual low state. The spectrum is characterized by the presence of strong He II and C IV emission lines as well as the normal single peaked Balmer emission lines. The radial temperature profile of the disk follows the expected T ~R^3/4 only in the outer parts and flattens off inside 0.5 times the white dwarf Roche lobe radius. The single peaked emission lines originate in a region above the plane of the disk, at the position of the hot spot.

We present an analysis of RXTE data of the X-ray transient XTE J1550-564. The source went through several states, which were divided into spectrally soft and hard states. These states showed up as distinct branches in the color-color diagram, forming a structure with a comb-like topology; the soft state branch forming the spine and the hard state branches forming the teeth. Variability was strongly correlated with the position on the branches. The broad band noise became stronger, and changed from power law like to band limited, as the spectrum became harder. Three types of QPOs were found: 1-18 Hz and 102-284 Hz QPOs on the hard branches, and 16-18 Hz QPOs on and near the soft branch. The frequencies of the high and low frequency QPOs on the hard branches were correlated with each other, and anti-correlated with spectral hardness. The changes in QPO frequency suggest that the inner disc radius only increases by a factor of 3-4 as the source changes from a soft to a hard state. Our results on XTE J1550-564 strongly favor a 2-dimensional description of black hole behavior, where the regions near the spine of the comb in the color-color diagram can be identified with the high state, and the teeth with transitions from the high state, via the intermediate state (which includes the very high state) to the low state, and back. The two physical parameters underlying this behavior vary to a large extent independently and could for example be the mass accretion rate through the disk and the size of a Comptonizing region.

We have searched the rising portion of type I X-ray bursts observed from the Rapid Burster with the Rossi X-ray Timing Explorer for the presence of periodicities. The 95 per cent confidence upper limit on the average root-mean-square variation of near coherent pulsations with a width of <1 Hz (in 60--2048 Hz) during the first second of the bursts is <8.8 per cent. We find a possible detection (>98 per cent significance) at 306.5 Hz.

We present time-resolved optical spectroscopy of the afterglows of the gamma-ray bursts GRB 990510 and GRB 990712. Through the identification of several absorption lines in the first epoch GRB 990510 spectrum, we determine the redshift for this burst at z >= 1.619. No clear emission lines are detected. The strength of the MgI feature is indicative of a dense environment, most likely the host galaxy of GRB 990510. Although the host is extremely faint (V >~ 28), the GRB afterglow allows us to probe its interstellar medium and - in principle - to measure its metallicity. The optical spectrum of GRB 990712 (whose host galaxy is the brightest of the known GRB hosts at cosmological redshifts), shows clear features both in emission and absorption, at a redshift of z = 0.4331 +/- 0.0004. On the basis of several line emission diagnostic diagrams, we conclude that the host galaxy of GRB 990712 is most likely an HII galaxy. We derive a unreddened [OII] star formation rate of 2.7 +/- 0.8 Msun/yr. Correcting for the measured extinction intrinsic to the host galaxy (A(V) = \gpm{3.4}{2.4}{1.7}), this value increases to \gpm{35}{178}{25} Msun/yr. The [OII] equivalent width, compared to that of field galaxies at z <= 1, also suggests that the host of GRB 990712 is vigorously forming stars. We employ the oxygen and Hbeta emission-line intensities to estimate the global oxygen abundance for the host of GRB 990712: log(O/H) = -3.7 +- 0.4, which is slightly below the lowest metallicity one finds in nearby spiral galaxies. For both GRBs we study the time evolution of the absorption lines, whose equivalent width might be expected to change with time if the burst resides in a dense compact medium. We find no evidence for a significant change in the MgII width.

The Soft Gamma Repeater SGR 1900+14 entered a remarkable phase of activity during the summer of 1998. This activity peaked on August 27, 1998 when a giant periodic gamma-ray flare resembling the famous March 5, 1979 event from SGR 0526-66 was recorded. Two days later (August 29), a strong, bright burst was detected with RXTE and BATSE. This event reveals several similarities to the giant flares of August 27 and March 5, and shows a number of unique features not previously seen in SGR bursts. Unlike typically short SGR bursts, this event features a 3.5 s burst peak that was preceded by an extended (~ 1 s) complex precursor, and followed by a long (~ 1000 s) periodic tail modulated at the 5.16 s stellar rotation period. Spectral analysis shows a striking distinction between the spectral behavior of the precursor, burst peak and extended tail. While the spectrum during the peak is uniform, a significant spectral evolution is detected in both the precursor and tail emissions. Temporal behavior shows a sharp rise (~ 9.8 ms) at the event onset and a rapid cutoff (~ 17 ms) at the end of the burst peak. The tail pulsations show a simple pulse profile consisting of one 5.16 s peak that did not evolve with time. The spectral and temporal signatures of this event imply that the precursor, main peak, and extended tail are produced by different physical mechanisms. We discuss these features and their implications in the context of the magnetar model. The energetics of the August 29 event, and its close proximity to the August 27 flare, suggest that it is an `aftershock' of the preceding giant flare. "P.S. This is an abbreviated version of the original abstract."

We report the detection of significant polarization in the optical afterglow of GRB990712 on three instances 0.44, 0.70 and 1.45 days after the gamma-ray burst, with (P, theta) being (2.9% +- 0.4%, 121.1 degr +- 3.5 degr), (1.2% +- 0.4%, 116.2 degr +- 10.1 degr) and (2.2% +- 0.7%, 139.2 degr +- 10.4 degr) respectively. The polarization is intrinsic to the afterglow. The degree of polarization is not constant, and smallest at the second measurement. The polarization angle does not vary significantly during these observations. We find that none of the existing models predict such polarization variations at constant polarization angle, and suggest ways in which these models might be modified to accommodate the observed behavior of this afterglow.

The Infrared Space Observatory observed the field of the $\gamma$--ray burst GRB 970508 with the CAM and PHT instruments on May 21 and 24, 1997 and with PHT in three filters in November 1997. A source at 60 $\mu$m (flux in May of $66\pm 10$ mJy) was detected near the position of the host galaxy of this $\gamma$--ray burst. The source was detected again in November 1997, at a marginally lower flux ($43\pm 13$ mJy). A Galactic cirrus origin and a stellar origin for the emission can be ruled out on the basis of the infrared colours. The marginal evidence for variability in the 60 $\mu$m flux between May and November is not sufficient to warrant interpretation of the source as transient fireball emission. However, the infrared colours are physically reasonable if attributed to conventional dust emission from a single blackbody source. The probability of detecting a 60 $\mu$m by chance in a PHT beam down to a detection limit of 50 mJy is $\sim 5\times 10^{-3}$. If the source is at the redshift of the host galaxy of the $\gamma$--ray burst the fluxes and upper limits at wavelengths from 12 $\mu$m to 170 $\mu$m indicate it is an ultraluminous infrared galaxy (L$_{\rm ir} \sim 2\times 10^{12}$ L$_{\sun}$). The star formation rate is estimated to be several hundred solar masses per year, depending significantly on model-dependent parameters. If this source is associated with the host galaxy of GRB 970508, progenitor models which associate GRBs with star-forming regions are favoured.

Using the exceptional long-term monitoring capabilities of the MACHO project, we present here the optical history of LMC X-2 for a continuous 6-yr period. These data were used to investigate the previously claimed periodicities for this source of 8.15 h and 12.54 d : we find upper amplitude limits of 0.10 mag and 0.09 mag, respectively.

The hard X-ray bursts observed during both major outbursts of the Bursting Pulsar (GRO J1744-28) show pulsations near the neutron star spin frequency with an enhanced amplitude relative to that of the persistent emission. Consistent with previous work, we find that the pulsations within bursts lag behind their expected arrival times based upon the persistent pulsar ephemeris. For an ensemble of 1293 bursts recorded with the Burst and Transient Source Experiment, the average burst pulse time delay is 61.0 +/- 0.8 ms in the 25 - 50 keV energy range and 72 +/- 5 ms in the 50 - 100 keV band. The residual time delay from 10 to 240 s following the start of the burst is 18.1 +/- 0.7 ms (25 - 50 keV). A significant correlation of the average burst time delay with burst peak flux is found. Our results are consistent with the model of the pulse time lags presented by Miller (1996).

We present statistics of SGR 1806-20 bursts, combining 290 events detected with RXTE/PCA, 111 events detected with BATSE and 134 events detected with ICE. We find that the fluence distribution of bursts observed with each instrument are well described by power laws with indices 1.43, 1.76 and 1.67, respectively. The distribution of time intervals between successive bursts from SGR 1806-20 is described by a lognormal function with a peak at 103 s. There is no correlation between the burst intensity and either the waiting times till the next burst or the time elapsed since the previous burst. In all these statistical properties, SGR 1806-20 bursts resemble a self-organized critical system, similar to earthquakes and solar flares. Our results thus support the hypothesis that the energy source for SGR bursts is crustquakes due to the evolving, strong magnetic field of the neutron star, rather than any accretion or nuclear power.

We measure simultaneously the properties of the energy spectra and the frequencies of the kilohertz quasi-periodic oscillations (QPOs) in fifteen low mass X-ray binaries covering a wide range of X-ray luminosities. In each source the QPO frequencies cover the same range of approximately 300 Hz to 1300 Hz, though the sources differ by two orders of magnitude in their X-ray luminosities (as measured from the unabsorbed 2-50 keV flux). So the X-ray luminosity does not uniquely determine the QPO frequency. This is difficult to understand since the evidence from individual sources indicates that the frequency and luminosity are very well correlated at least over short timescales. Perhaps beaming effects or bolometric corrections change the observed luminosities, or perhaps part of the energy in mass accretion is used to power outflows reducing the energy emitted in X-rays. It is also possible that the parameters of a QPO model are tuned in such a way that the same range of frequencies appears in all sources. Different modes of accretion may be involved for example (disk and radial) or multiple parameters may conspire to yield the same frequencies.

We present an analysis of ~390 ksec of data of the Z source GX 340+0 taken during 24 observations with the Rossi X-ray Timing Explorer satellite. We report the discovery of a new broad component in the power spectra. The frequency of this component varied between 9 and 14 Hz, and remained close to half that of the horizontal branch quasi-periodic oscillations (HBO). Its rms amplitude was consistent with being constant around ~5%, while its FWHM increased with frequency from 7 to 18 Hz. If this sub-HBO component is the fundamental frequency, then the HBO and its second harmonic are the second and fourth harmonic component, while the third harmonic was not detected. This is similar to what was recently found for the black hole candidate XTE J1550-564. The profiles of both the horizontal- and the normal branch quasi-periodic oscillation peaks were asymmetric when they were strongest. We describe this in terms of a shoulder component at the high frequency side of the quasi-periodic oscillation peak, whose rms amplitudes were approximately constant at ~4% and \~3%, respectively. The peak separation between the twin kHz quasi-periodic oscillations was consistent with being constant at 339+-8 Hz but a trend similar to that seen in, e.g. Sco X-1 could not be excluded. We discuss our results within the framework of the various models which have been proposed for the kHz QPOs and low frequency peaks.

SN 1998bw holds the record for the most energetic Type Ic explosion, one of the brightest radio supenovae and probably the first supernova associated with a gamma-ray burst. In this paper we present spectral observations of SN 1998bw observed in a cooperative monitoring campaign using the AAT, UKST and the SSO 2.3-m telescope. We investigate the evolution of the spectrum between 7 and 94 days after V_band maximum in comparison to well-studied examples of Type Ic SNe in order to quantify the unusual properties of this supernova event. Though the early spectra differ greatly from the observations of classical Ic SNe, we find that the evolution from the photospheric to the nebular phases is slow but otherwise typical. The spectra differ predominantly in the extensive line blending and blanketing which has been attributed to the high velocity of the ejecta. We find that by day 19, the absorption line minima blueshifts are 10% - 50% higher than other SNe and on day 94 emission lines are 45% broader, as expected if the progenitor had a massive envelope. However, it is difficult to explain the extent of line blanketing entirely by line broadening, and we argue that additional contribution from other species is present, indicating unusual relative abundances or physical conditions in the envelope.

We have obtained intermediate resolution optical spectra of the black-hole candidate Nova Sco 1994 in June 1996, when the source was in an X-ray/optical active state (R~15.05). We measure the radial velocity curve of the secondary star and obtain a semi-amplitude of 279+/-10 km/s; a value which is 30 per cent larger than the value obtained when the source is in quiescence. Our large value for K_2 is consistent with 60 +9,-7 per cent of the secondary star's surface being heated; compared to 35 per cent, which is what one would expect if only the inner face of the secondary star were irradiated. Effects such as irradiation-induced flows on the secondary star may be important in explaining the observed large value for K_2.

We analyzed 163 observations of the microquasar GRS 1915+105 made with the Rossi X-ray Timing Explorer (RXTE) in the period 1996-1997. For each observation, we produced light curves and color-color diagrams. We classified the observations in 12 separate classes, based on their count rate and color characteristics. From the analysis of these classes, we reduced the variability of the source to transitions between three basic states: a hard state corresponding to the non-observability of the innermost parts of the accretion disk, and two softer states with a fully observable disk. These two soft states represent different temperatures of the accretion disk, related to different local values of the accretion rate. The transitions between these states can be extremely fast. The source moves between these three states following certain patterns and avoiding others, giving rise to a relatively large but limited number of variability classes. These results are the first step towards a linking of the properties of this exceptional source with standard black-hole systems and with accretion disk models.

We analyze 426 observations of the bursting pulsar GRO J1744-28 by Ulysses and BATSE. Triangulating each burst, and statistically combining the triangulation annuli, we obtain a 3 sigma error ellipse whose area is 532 sq. arcsec. The accuracy of this statistical method has been independently verified with observations of the soft gamma repeater SGR1900+14. The ellipse is fully contained within the 1 ' radius ASCA error circle of the soft X-ray counterpart, and partially overlaps the 10 " radius ROSAT error circle of a source which may also be the soft X-ray counterpart. A variable source which has been proposed as a possible IR counterpart lies at the edge of the 3 sigma error ellipse, making it unlikely from a purely statistical point of view to be associated with the bursting pulsar.

We report on the evolution of the timing of 4U1630-47 during its 1998 outburst using data obtained with the Rossi X-ray Timing Explorer (RXTE). The count rate and position in hardness-intensity, color-color diagrams and simple spectral fits are used to track the concurrent spectral changes. The source showed seven distinct types of timing behavior, most of which show differences with the canonical black hole spectral/timing states. In marked contrast to previous outbursts, we find quasi-periodic oscillation (QPO) signals during nearly all stages of the outburst with frequencies between 0.06Hz and 14Hz and a remarkable variety of other characteristics. In particular we find large (up to 23% rms) amplitude QPO on the early rise. Later, slow 0.1Hz semi- regular short (~ 5 sec), 9 to 16% deep dips dominate the light curve. At this time there are two QPOs, one stable near 13.5Hz and the other whose frequency drops from 6--8Hz to ~4.5 Hz during the dips. BeppoSAX observations during the very late declining phase show 4U1630-47 in a low state.

We have identified a likely radio counterpart to the low-mass X-ray binary MXB 1730-335 (the Rapid Burster). The counterpart has shown 8.4 GHz radio on/off behavior correlated with the X-ray on/off behavior as observed by the RXTE/ASM during six VLA observations. The probability of an unrelated, randomly varying background source duplicating this behavior is 1-3% depending on the correlation time scale. The location of the radio source is RA 17h 33m 24.61s; Dec -33d 23' 19.8" (J2000), +/- 0.1". We do not detect 8.4 GHz radio emission coincident with type II (accretion-driven) X-ray bursts. The ratio of radio to X-ray emission during such bursts is constrained to be below the ratio observed during X-ray persistent emission at the 2.9-sigma level. Synchrotron bubble models of the radio emission can provide a reasonable fit to the full data set, collected over several outbursts, assuming that the radio evolution is the same from outburst to outburst, but given the physical constraints the emission is more likely to be due to ~hour-long radio flares such as have been observed from the X-ray binary GRS 1915+105.

We have discovered a ~1 Hz quasi-periodic oscillation (QPO) in the persistent X-ray emission and during type I X-ray bursts of the globular cluster source, dipper and low-mass X-ray binary (LMXB) 4U 1746-37. The QPO properties resemble those of QPOs found recently in the LMXB dippers 4U 1323-62, and EXO 0748-676, which makes 4U 1746-37 the third source known to exhibit this type of QPOs. We present evidence for X-ray spectral changes in this source similar to those observed in LMXBs referred to as atoll sources. We detect two states, a low intensity and spectrally hard state, and a higher intensity and spectrally soft state. This may explain the different spectral characteristics reported for 4U 1746-37 earlier. The high intensity state resembles the banana branch state of atoll sources. The QPOs are only seen in the low intensity state, and are absent when the source is in the banana branch. This strongly suggests that either the accretion disk or an extended central source change shape between the low intensity state and the banana branch. Twelve bursts were detected, of which 5 took place while the source was on the banana branch and 7 when the source was in the low intensity state. The bursts occurring on the banana branch had an e-folding time ~3 times longer than those which occurred in the low intensity state. Whereas previously detected dips showed only a decrease in count rate of ~15%, we found in one observation a dip in which the count rate dropped from ~200 counts per second to ~20 counts per second. This dip lasted only ~250 seconds, during which clear spectral hardening occured. This is the first time strong evidence for spectral changes during a dip are reported for this source.

We study the statistics of soft gamma repeater (SGR) bursts, using a data base of 187 events detected with BATSE and 837 events detected with RXTE PCA, all from SGR 1900+14 during its 1998-1999 active phase. We find that the fluence or energy distribution of bursts is consistent with a power law of index 1.66, over 4 orders of magnitude. This scale-free distribution resembles the Gutenberg-Richter Law for earthquakes, and gives evidence for self-organized criticality in SGRs. The distribution of time intervals between successive bursts from SGR 1900+14 is consistent with a log-normal distribution. There is no correlation between burst intensity and the waiting times till the next burst, but there is some evidence for a correlation between burst intensity and the time elapsed since the previous burst. We also find a correlation between the duration and the energy of the bursts, but with significant scatter. In all these statistical properties, SGR bursts resemble earthquakes and solar flares more closely than they resemble any known accretion-powered or nuclear-powered phenomena. Thus our analysis lends support to the hypothesis that the energy source for SGR bursts is internal to the neutron star, and plausibly magnetic.

We present evidence for burst emission from SGR 1900+14 with a power-law high energy spectrum extending beyond 500 keV. Unlike previous detections of high energy photons during bursts from SGRs, these emissions are not associated with high-luminosity burst intervals. Not only is the emission hard, but the spectra are better fit by Band's GRB function rather than by the traditional optically-thin thermal bremsstrahlung model. We find that the spectral evolution within these hard events obeys a hardness/intensity anti-correlation. Temporally, these events are distinct from typical SGR burst emissions in that they are longer (~ 1 s) and have relatively smooth profiles. Despite a difference in peak luminosity of > 1E+11 between these bursts from SGR 1900+14 and cosmological GRBs, there are striking temporal and spectral similarities between the two kinds of bursts, aside from spectral evolution. We outline an interpretation of these events in the context of the magnetar model.

We present a 2 - 10 keV ASCA observation of the field around the soft gamma repeater SGR1627-41. A quiescent X-ray source was detected in this observation whose position was consistent both with that of a recently discovered BeppoSAX X-ray source and with the Interplanetary Network localization for this SGR. In 2 - 10 keV X-rays, the spectrum of the X-ray source may be fit equally well by a power law, blackbody, or bremsstrahlung function, with unabsorbed flux ~5 x 10^-12 erg cm^-2 s^-1. We do not confirm a continuation of a fading trend in the flux, and we find no evidence for periodicity, both noted in the earlier BeppoSAXobservations.

We present BVRI photometry and spectrophotometry of GRB990510 obtained with the ESO VLT/Antu telescope during the late decline phase. Between days 8 and 29 after the burst, the afterglow faded from R=24.2 to ~26.4. The spectral flux distribution and the light curve support the interpretation of the afterglow as synchrotron emission from a jet. The light curve is consistent with the optical transient alone but an underlying SN with maximum brightness R>27.4 or a galaxy with R>27.6 (3-sigma upper limits) cannot be ruled out. To a 5-sigma detection threshold of R=26.1, no galaxy is found within 6'' of the transient. A very blue V~24.5 extended object which may qualify as a starburst galaxy is located 12'' SE, but at unknown redshift.

We report on two observations of a soft bursting source in 1997 June, whose time histories and energy spectra are consistent with those of the soft gamma repeaters. The source can only be localized to an ~3.8 degree long error box in the direction of the Galactic center, whose area is ~ 80 sq. arcmin. The location of the source, while not consistent with that of any of the four known soft repeaters, is consistent with those of several known and possible supernova remnants.

In this Letter, we present the first evidence in the BATSE data for a prompt high-energy (25-300 keV) afterglow component from a gamma-ray burst (GRB), GRB980923. The event consists of rapid variabilty lasting ~40 s followed by a smooth power law emission tail lasting ~400 s. An abrupt change in spectral shape is found when the tail becomes noticeable. Our analysis reveals that the spectral evolution in the tail of the burst mimics that of a cooling synchrotron spectrum, similar to the spectral evolution of the low-energy afterglows for GRBs. This evidence for a separate emission component is consistent with the internal-external shock scenario in the relativistic fireball picture. In particular, it illustrates that the external shocks can be generated during the gamma-ray emission phase, as in the case of GRB990123.

We consider the physical implications of the rapid spindown of Soft Gamma Repeater 1900+14, and of the apparent "braking glitch", \Delta P/P = l x 10^-4, that was concurrent with the Aug. 27th giant flare. A radiation-hydrodynamical outflow associated with the flare could impart the required torque, but only if the dipole magnetic field is stronger than ~ 10^14 G and the outflow lasts longer and/or is more energetic than the observed X-ray flare. A positive period increment is also a natural consequence of a gradual, plastic deformation of the neutron star crust by an intense magnetic field, which forces the neutron superfluid to rotate more slowly than the crust. Sudden unpinning of the neutron vortex lines during the August 27th event would then induce a glitch opposite in sign to those observed in young pulsars, but of a much larger magnitude as a result of the slower rotation. The change in the persistent X-ray lightcurve following the August 27 event is ascribed to continued particle heating in the active region of that outburst. The enhanced X-ray output can be powered by a steady current flowing through the magnetosphere, induced by the twisting motion of the crust. The long term rate of spindown appears to be accelerated with respect to a simple magnetic dipole torque. Accelerated spindown of a seismically-active magnetar will occur when its persistent output of Alfven waves and particles exceeds its spindown luminosity. We suggest that SGRs experience some episodes of relative inactivity, with diminished spindown rates, and that such inactive magnetars are observed as Anomalous X-ray Pulsars (AXPs). The rapid reappearence of persistent X-ray emission following August 27 flare gives evidence against accretion-powered models.

We present B-, V-, R_c-, I_c-, J-, H-, K- and K'-band observations of the optical transient (OT) associated with GRB970228, based on a reanalysis of previously used images and unpublished data. In order to minimize calibration differences we have collected and analyzed most of the photometry and consistently determined the magnitude of the OT relative to a set of secondary field stars. We confirm our earlier finding that the early decay of the light curves (before March 6, 1997) was faster than that at intermediate times (between March 6 and April 7, 1997). At late times the light curves resume a fast decay (after April 7, 1997). The early-time observations of GRB970228 are consistent with relativistic blast-wave models but the intermediate- and late-time observations are hard to understand in this framework. The observations are well explained by an initial power law decay with index -1.73 +0.09 -0.12 modified at later times by a type-I_c supernova light curve. Together with the evidence for GRB980326 and GRB980425 this gives further support for the idea that at least some GRBs are associated with a possibly rare type of supernova.

We have analyzed Rossi X-ray Timing Explorer Proportional Counter Array observations of the pulsed emission from SGR 1900+14 during September 1996, June - October 1998, and early 1999. Using these measurements and results reported elsewhere, we construct a period history of this source for 2.5 years. We find significant deviations from a steady spin-down trend during quiescence and the burst active interval. Burst and Transient Source Experiment observations of the burst emission are presented and correlations between the burst activity and spin-down rate of SGR 1900+14 are discussed. We find an 80 day interval during the summer of 1998 when the average spin-down rate is larger than the rate elsewhere by a factor ~ 2.3. This enhanced spin-down may be the result of a discontinuous spin-down event or ``braking glitch'' at the time of the giant flare on 27 August 1998. Furthermore, we find a large discrepancy between the pulsar period and average spin-down rate in X-rays as compared to radio observations for December 1998 and January 1999.

We present the results of the analysis of RXTE, BATSE and optical/infrared data of the 1998 outburst of the X-ray transient system XTE J0421+560 (CI Cam). The X-ray outburst shows a very fast decay (initial e-folding time ~0.5 days, slowing down to ~2.3 days). The X-ray spectrum in the 2-25 keV band is complex, softening considerably during decay and with strongly variable intrinsic absorption. A strong iron emission line is observed. No fast time variability is detected (<0.5 % rms in the 1-4096 Hz band at the outburst peak). The analysis of the optical/IR data suggests that the secondary is a B[e] star surrounded by cool dust and places the system at a distance of >~ 2 kpc. At this distance the peak 2-25 keV luminosity is ~4 x 10^37 erg/s. We compare the properties of this peculiar system with those of the Be/NS LMC transient A 0538-66 and suggest that CI Cam is of similar nature. The presence of strong radio emission during outburst indicates that the compact object is likely to be a black hole or a weakly magnetized neutron star.

We apply a statistical method to derive very precise locations for soft gamma repeaters using data from the interplanetary network. We demonstrate the validity of the method by deriving a 600 arcsec^2 error ellipse for SGR1900+14 whose center agrees well with the VLA source position. We then apply it to SGR1806-20, for which we obtain a 230 arcsec^2 error ellipse, the smallest burst error box to date. We find that the most likely position of the source has a small but significant displacement from that of the non-thermal core of the radio supernova remnant G10.0-0.3, which was previously thought to be the position of the repeater. We propose a different model to explain the changing supernova remnant morphology and the positions of the luminous blue variable and the bursting source.

Following a BeppoSAX alert (Piro 1999a) and the discovery of the OT at SAAO (Vreeswijk et al. 1999a), we observed GRB 990510 with the FORS instrument on ESO's VLT Unit 1 (`Antu'). The burst is unremarkable in gamma rays, but in optical is the first one to show good evidence for jet-like outflow (Stanek et al. 1999, Harrison et al. 1999). We report the detection of significant linear polarization in the afterglow: it is (1.6 +/- 0.2)% 0.86 days after trigger, and after 1.81 days is consistent with that same value, but much more uncertain. The polarization angle is constant on a time scale of hours, and may be constant over one day. We conclude that the polarization is intrinsic to the source and due to the synchrotron nature of the emission, and discuss the random and ordered field geometries that may be responsible for it.

We report on observations of the fading optical counterpart of the gamma-ray burst GRB 970228, made with the Hubble Space Telescope STIS CCD approximately six months after outburst and with the HST/NICMOS and Keck/NIRC approximately one year after outburst. The unresolved counterpart is detected by STIS at V=28.0 +/- 0.25, consistent with a continued power-law decline with exponent -1.14 +/- 0.05. The counterpart is located within, but near the edge of, a faint extended source with diameter ~0."8 and integrated magnitude V=25.8 +/- 0.25. A reanalysis of HST and NTT observations performed shortly after the burst shows no evidence of proper motion of the point source or fading of the extended emission. Only the extended source is visible in the NICMOS images with a magnitude of H=23.3 +/- 0.1. The Keck observations find K = 22.8 +/- 0.3. Several distinct and independent means of deriving the foreground extinction in the direction of GRB 970228 all agree with A_V = 0.75 +/- 0.2. After adjusting for Galactic extinction, we find that the size of the observed extended emission is consistent with that of galaxies of comparable magnitude found in the Hubble Deep Field (HDF) and other deep HST images. Only 2% of the sky is covered by galaxies of similar or greater surface brightness; therefore the extended source is almost certainly the host galaxy. Additionally, we find that the extinction-corrected V - H and V - K colors of the host are as blue as any galaxy of comparable or brighter magnitude in the HDF. Taken in concert with recent observations of GRB 970508, GRB 971214, and GRB 980703 our work suggests that all four GRBs with spectroscopic identification or deep multicolor broad-band imaging of the host lie in rapidly star-forming galaxies.

We report on HST and Palomar optical images of the field of GRB 990123, obtained on 8 and 9 February 1999. We find that the optical transient (OT) associated with GRB 990123 is located on an irregular galaxy, with magnitude V=24.20 +/- 0.15. The strong metal absorption lines seen in the spectrum of the OT, along with the low probability of a chance superposition, lead us to conclude that this galaxy is the host of the GRB. The OT is projected within the ~1'' visible stellar field of the host, nearer the edge than the center. We cannot, on this basis, rule out the galactic nucleus as the site of the GRB, since the unusual morphology of the host may be the result of an ongoing galactic merger, but our demonstration that this host galaxy has extremely blue optical to infrared colors more strongly supports an association between GRBs and star formation. We find that the OT magnitude on 1999 Feb 9.05, V = 25.45 +/- 0.15, is about 1.5 mag fainter than expected from extrapolation of the decay rate found in earlier observations. A detailed analysis of the OT light curve suggests that its fading has gone through three distinct phases: an early rapid decline (f_{nu} \propto t^{-1.6} for t < 0.1 days), a slower intermediate decline power-law decay (f_{nu} \propto t^{-1.1} for 0.1 < t < 2 days), and then a more rapid decay (at least as steep as (f_{\nu} \propto t^{-1.8} for t > 2 days). The break to steeper slope at late times may provide evidence that the optical emission from this GRB was highly beamed.

We describe the evolutionary progression of an outburst of the Rapid Burster. Four outbursts have been observed with the Rossi X-Ray Timing Explorer between February 1996 and May 1998, and our observations are consistent with a standard evolution over the course of each. An outburst can be divided into two distinct phases: Phase I is dominated by type I bursts, with a strong persistent emission component; it lasts for 15-20 days. Phase II is characterized by type II bursts, which occur in a variety of patterns. The light curves of time-averaged luminosity for the outbursts show some evidence for reflares, similar to those seen in soft X-ray transients. The average recurrence time for Rapid Burster outbursts during this period has been 218 days, in contrast with an average ~180 day recurrence period observed during 1976-1983.

We report on X-ray, optical and infrared follow-up observations of GRB 980703. We detect a previously unknown X-ray source in the GRB error box; assuming a power law decline we find for its decay index alpha < -0.91 (3-sigma). We invoke host galaxy extinction to match the observed spectral slope with the slope expected from `fireball' models. We find no evidence for a spectral break in the infrared to X-ray spectral range on 1998 July 4.4, and determine a lower limit of the cooling break frequency: nu_c > 1.3 x 10^{17} Hz. For this epoch we obtain an extinction of A_V = 1.50 +/- 0.11. From the X-ray data we estimate the optical extinction to be A_V = 20.2 +12.3 -7.3, inconsistent with the former value. Our optical spectra confirm the redshift of z = 0.966 found by Djorgovski et al. (1998). We compare the afterglow of GRB 980703 with that of GRB 970508 and find that the fraction of the energy in the magnetic field, epsilon_B < 6x10^{-5}, is much lower in the case of GRB 980703, which is a consequence of the high frequency of the cooling break.

The 8.7s X-ray pulsar 4U 0142+61 was monitored by BeppoSAX between January 1997 and February 1998. This source belongs to the rapidly growing class of "anomalous" X-ray pulsars (AXPs) which have pulse periods in the 6-12s range and no plausible optical, infrared or radio counterparts. The BeppoSAX periods measurements show that 4U 0142+61 continues its spin-down at a nearly constant rate of about 2 times 10^{-12}s s^{-1}. The 0.5-10 keV pulse shape is double peaked. The phase-averaged spectrum can be described by a steep absorbed power-law (Gamma about 4) plus a blackbody with kT = 0.4 keV. 4U 0142+61 was also detected serendipitously in a March 1996 RossiXTE observation pointed towards the nearby 1455s X-ray pulsar RX J0146.9+6121. The timing analysis results are reported for this observation and the spin history of 4U 0142+61 since 1979 is discussed.

We have used the ROSAT High Resolution Imager to search for quiescent X-ray counterparts to four gamma-ray bursts which were localized to small (< 10 sq. arcmin.) error boxes with the Interplanetary Network. The observations took place years after the bursts, and the effective exposure times for each target varied from ~16 - 23 ks. We have not found any X-ray sources inside any of the error boxes. The 0.1 - 2.4 keV 3 sigma flux upper limits range from around 5 x 10^-14 erg cm^-2 s^-1 to 6 x 10^-13 erg cm^-2 s^-1 depending on the burst and the assumed shape of the quiescent spectrum. We consider four types of X-ray emitting galaxies (normal, AGN, faint, and star-forming) and use the flux upper limits to constrain their redshifts. We then use the GRB fluences to constrain the total energies of the bursts.

We present Ulysses, KONUS-WIND, and BATSE observations of bursts from a new soft gamma repeater which was active in 1998 June and July. Triangulation of the bursts results in a ~ 1.8 degree by 16 '' error box whose area is ~ 7.6 arcminutes^2, which contains the Galactic supernova remnant G337.0-0.1. This error box intersects the position of a BeppoSAX X-ray source which is also consistent with the position of G337.0-0.1 (Woods et al. 1999), and is thought to be the quiescent counterpart to the repeater. If so, the resulting error box is ~ 2 ' by 16 '' and has an area of ~ 0.6 arcminutes^2. The error box location within the supernova remnant suggests that the neutron star has a transverse velocity of ~ 200 - 2000 km/s.

We report the discovery of a new soft gamma repeater (SGR), SGR 1627-41, and present BATSE observations of the burst emission and BeppoSAX NFI observations of the probable persistent X-ray counterpart to this SGR. All but one burst spectrum are well fit by an optically thin thermal bremsstrahlung (OTTB) model with kT values between 25 and 35 keV. The spectrum of the X-ray counterpart, SAX J1635.8-4736, is similar to that of other persistent SGR X-ray counterparts. We find weak evidence for a periodic signal at 6.41 s in the light curve for this source. Like other SGRs, this source appears to be associated with a young supernova remnant G337.0-0.1. Based upon the peak luminosities of bursts observed from this SGR, we find a lower limit on the dipole magnetic field of the neutron star B_dipole > 5 * 10^14 Gauss.

GRB 990123 was the first burst from which simultaneous optical, X-ray and gamma-ray emission was detected; its afterglow has been followed by an extensive set of radio, optical and X-ray observations. We have studied the gamma-ray burst itself as observed by the CGRO detectors. We find that gamma-ray fluxes are not correlated with the simultaneous optical observations, and the gamma-ray spectra cannot be extrapolated simply to the optical fluxes. The burst is well fit by the standard four-parameter GRB function, with the exception that excess emission compared to this function is observed below ~15 keV during some time intervals. The burst is characterized by the typical hard-to-soft and hardness-intensity correlation spectral evolution patterns. The energy of the peak of the nu f_nu spectrum, E_p, reaches an unusually high value during the first intensity spike, 1470 +/- 110 keV, and then falls to \~300 keV during the tail of the burst. The high-energy spectrum above ~MeV is consistent with a power law with a photon index of about -3. By fluence, GRB 990123 is brighter than all but 0.4% of the GRBs observed with BATSE, clearly placing it on the -3/2 power-law portion of the intensity distribution. However, the redshift measured for the afterglow is inconsistent with the Euclidean interpretation of the -3/2 power-law. Using the redshift value of >= 1.61 and assuming isotropic emission, the gamma-ray fluence exceeds 10E54 ergs.

We report on observations of the field of GRB~970508 made in early August 1998, 454 days after outburst, with the STIS CCD camera onboard the Hubble Space Telescope. The images, taken in open filter (50CCD) mode, clearly reveal the presence of a galaxy which was obscured in earlier (June 1997) HST images by emission from the optical transient (OT). The galaxy is regular in shape: after correcting for the HST/STIS PSF, it is well-fitted by an exponential disk with a scale length of 0."046 +/- 0."006 and an ellipticity of 0.70 +/- 0.07. All observations are marginally consistent with a continuous decline in OT emission as t^{-1.3} beginning two days after outburst; however, we find no direct evidence in the image for emission from the OT, and the surface brightness profile of the galaxy is most regular if we assume that the OT emission is negligible, suggesting that the OT may have faded more rapidly at late times than is predicted by the power-law decay. Due to the wide bandwidth of the STIS clear mode, the estimated magnitude of the galaxy is dependent on the galaxy spectrum that is assumed. Using colors obtained from late-time ground-based observations to constrain the spectrum, we find V = 25.4 +/- 0.15, a few tenths of a magnitude brighter than earlier ground-based estimates that were obtained by observing the total light of the galaxy and the OT and then subtracting the estimated OT brightness assuming it fades as a single power-law. This again suggests that the OT may have faded faster at late time than the power-law predicts. The position of the OT agrees with that of the isophotal center of the galaxy to 0."01. This remarkable agreement raises the possibility that the GRB may have been associated with either an active galactic nucleus or a nuclear starburst.

We report the discovery of a quasi-periodic oscillation (QPO) in data obtained with the Rossi X-ray Timing Explorer of the dipping and eclipsing low-mass X-ray binary EXO 0748-676. The QPO had a frequency between 0.58 and 2.44 Hz changing on time scales of a few days, an rms amplitude between 8% and 12%, and was detected in the persistent emission, during dips and during type I X-ray bursts. During one observation, when the count rate was a factor 2 to 3 higher than otherwise, the QPO was not detected. The strength of the QPO did not significantly depend on photon energy, and is consistent with being the same in the persistent emission, both during and outside the dips, and during type I X-ray bursts. Frequency shifts were observed during three of the four X-ray bursts. We argue that the QPO is produced by the same mechanism as the QPO recently found by Jonker et al. (1999) in 4U 1323-62. Although the exact mechanism is not clear, it is most likely related to the high inclination of both systems. An orbiting structure in the accretion disc that modulates the radiation from the central source seems the most promising mechanism.

We study the horizontal branch oscillations (HBO) and the two simultaneous kilohertz QPOs comparing their properties in five Z sources with the predictions of the magnetospheric beat-frequency and Lense-Thirring precession models. We find that the variation of the HBO frequency with accretion rate predicted by the magnetospheric beat-frequency model for a purely dipolar stellar magnetic field and a radiation-pressure-dominated inner accretion disk agrees well with the observed variation. This model implies that the neutron stars in the Z sources are near magnetic spin equilibrium, that their magnetic field strengths are ~10^9-10^10 G, and that the critical fastness parameter for these sources is >0.8. If the frequency of the upper kilohertz QPO is an orbital frequency in the accretion disk, the magnetospheric beat-frequency model requires that a small fraction of the gas in the disk does not couple strongly to the stellar magnetic field at 3-4 stellar radii but instead drifts slowly inward in nearly circular orbits until it is within a few kilometers of the neutron star surface. The Lense-Thirring precession model is consistent with the observed magnitudes of the HBO frequencies only if the moments of inertia of the neutron stars in the Z sources are 4-5 times larger than the largest values predicted by realistic neutron-star equations of state. The trend of the correlation between the HBO and upper kHz QPO frequencies is not consistent with this model. We argue that the change in the slope of the correlation between the frequency of the horizontal branch oscillation and the frequency of the upper kilohertz QPO, when the latter is greater than 850 Hz, is directly related to the varying frequency separation of the kilohertz QPOs.

Gamma-ray bursts (GRBs) are thought to result from the interaction of an extremely relativistic outflow interacting with a small amount of material surrounding the site of the explosion. Multi-wavelength observations covering the gamma-ray to radio wavebands allow investigations of this `fireball' model. On 23 January 1999 optical emission was detected while the gamma-ray burst was still underway. Here we report the results of gamma-ray, optical/infra-red, sub-mm, mm and radio observations of this burst and its afterglow, which indicate that the prompt and afterglow emissions from GRB 990123 are associated with three distinct regions in the fireball. The afterglow one day after the burst has a much lower peak frequency than those of previous bursts; this explains the short-lived nature of the radio emission, which is not expected to reappear. We suggest that such differences reflect variations in the magnetic-field strengths in the afterglow emitting regions.

Recent rapid advances in the observations of Gamma Ray Bursts (GRBs) lend much confidence to the cosmological fireball (FB) model: there is now compelling evidence that the radiation is emitted by a relativistic shock where a high speed upstream flow terminates. The question concerning what generates such an outflow is central to our search for the ultimate trigger mechanism. A key requirement, not well explained by current theories, is that the flow must have high entropy-to-baryon ratio. In this Letter we point out that a quantum discharge induced by the radiation of the initial FB may be the explanation. The effect is likely to be relevant, because the FB energy density inferred from GRB data is large enough that the radiation pressure leads to the formation of a surface electric field which is unstable to pair creation. Under suitable conditions, such as those of a supernova core, the discharge can convert a substantial fraction of the FB energy into surface pairs. This pair plasma is not contaminated by FB baryons because it is formed outside the FB. We demonstrate that the pairs can then develop relativistic bulk expansion, reaching a maximum speed that meets the constraints required to form a GRB.

We have obtained high and intermediate resolution optical spectra of the black-hole candidate Nova Sco 1994 in May/June 1998, when the source was in complete (X-ray) quiescence. We measure the radial velocity curve of the secondary star and obtain a semi-amplitude of K_2=215.5+/-2.4 km/s, which is 6 per cent lower than the only previously determined value. This new value for K_2 thus reduces the binary mass function to f(M) = 2.73+/-0.09 Mo. Using only the high resolution spectra we constrain the rotational broadening of the secondary star, vsini, to lie in the range 82.9-94.9 km/s (95 per cent confidence) and thus constrain the binary mass ratio to lie in the range 0.337--0.436 (95 per cent confidence). We can also combine our results with published limits for the binary inclination to constrain the mass of the compact object and secondary star to the ranges 5.5 -- 7.9 and 1.7 -- 3.3 Mo respectively (95 per cent confidence). Finally, we report on the detection of the Lithium resonance line at 6707.8 A, with an equivalent width of 55+/-8 mA.

We have used the Fourier cross spectra of GRO J1719-24, as obtained with BATSE, to estimate the phase lags between the X-ray flux variations in the 20--50 and 50--100 keV energy bands as a function of Fourier frequency in the interval 0.002--0.488 Hz. Our analysis covers the entire ~80 day X-ray outburst of this black-hole candidate, following the first X-ray detection on 1993 September 25. The X-ray variations in the 50--100 keV band lag those in the 20--50 keV energy band by an approximately constant phase difference of 0.072 +/- 0.010 rad in the frequency interval 0.02--0.20 Hz. The peak phase lags in the interval 0.02--0.20 Hz are about twice those of Cyg X-1 and GRO J0422+32.These results are consistent with models for Comptonization regions composed of extended non-uniform clouds around the central source.

We report the discovery of twin kHz quasi-periodic oscillations (QPOs) in the persistent X-ray flux of the new X-ray transient XTE J2123-058. We detected the QPOs in data taken with the proportional counter array on board the Rossi X-ray Timing Explorer on 1998 July 14, when the source count rate was at about 50% of its outburst peak level. The frequencies of the QPOs were 853+/-4 Hz and 1129+/-8 Hz. The peaks had widths of 32 (+11/-8) Hz and 50 (+17/-13) Hz and the rms amplitudes were 6.1% (+0.7/-0.6) and 6.5% (+0.9/-0.8), respectively. The QPO frequencies increased marginally with count rate and the amplitudes showed a small increase with photon energy. On the basis of the appearance in the color-color and hardness-intensity diagrams, the low-frequency power spectrum, and the strength of the kHz QPOs, the system can be classified as an atoll source. The source showed five type I X-ray bursts. The distance to the source is estimated to be about 10 (5-15) kpc, which combined with the high Galactic latitude of 36.2 degrees indicates the source is a located in the Galactic halo.

We have recently completed a search of 6 years of archival BATSE data for gamma-ray bursts (GRBs) that were too faint to activate the real-time burst detection system running onboard the spacecraft. These "non-triggered" bursts can be combined with the "triggered" bursts detected onboard to produce a GRB intensity distribution that reaches peak fluxes a factor of 2 lower than could be studied previously. The value of the V/Vmax statistic (in Euclidean space) for the bursts we detect is 0.177 +/- 0.006. This surprisingly low value is obtained because we detected very few bursts on the 4.096 s and 8.192 s time scales (where most bursts have their highest signal-to-noise ratio) that were not already detected on the 1.024 s time scale. If allowance is made for a power-law distribution of intrinsic peak luminosities, the extended peak flux distribution is consistent with models in which the redshift distribution of the gamma-ray burst rate approximately traces the star formation history of the Universe. We argue that this class of models is preferred over those in which the burst rate is independent of redshift. We use the peak flux distribution to derive a limit of 10% (99% confidence) on the fraction of the total burst rate that could be contributed by a spatially homogeneous (in Euclidean space) subpopulation of burst sources, such as type Ib/c supernovae. These results lend support to the conclusions of previous studies predicting that relatively few faint "classical" GRBs will be found below the BATSE onboard detection threshold.

We have obtained high resolution optical spectroscopy of the accretion disc corona source 2S0921--630 (=V395 Car) that constrains the spectral type of the companion star to be most probably K0III. At 6500A the secondary star is found to contribute ~25% to the observed flux. The absorption line spectrum of the K0III companion star is broadened (compared to template stars) by a projected rotational velocity of 65+/-9 km/s (1 sigma).

Using the Rossi X-ray Timing Explorer we have measured lags of the 9 to 33 keV photons relative to the 2 to 9 keV photons in the timing noise between 0.01 and 100 Hz in the accreting neutron stars 4U 0614+091 and 4U 1705-44. We performed similar measurements on the accreting black hole candidates Cyg X-1 and GX 339-4 as a comparison. During the observations these sources were all in low (hard) states. We find phase lags of between 0.03 and 0.2 radians in all these sources, with a variation in frequency much less than expected for a lag constant in time. We also measure a coherence consistent with unity in all sources. As already noted for the black hole candidates, these data are inconsistent with simple Comptonization models invoking a constant time delay. Comptonization in a non-uniform medium can perhaps explain the lags. However, the magnitudes of the lags imply that the hot electron gas extends to more than 1000 Schwarzschild radii. This may constitute an energy problem. We argue that while a large hot cloud is possible for black holes which may hide some of their accretion energy in advection, such a distribution may not be possible neutron stars where all the accretion energy is eventually released at the neutron star surface. This casts doubt on the Comptonization model, though the energy problem may be resolved for example by a wind from the inner disk.

The soft-gamma repeater SGR 1900+14 became active again on June 1998 after a long period of quiescence; it remained at a low state of activity until August 1998, when it emitted a series of extraordinarily intense outbursts. We have observed the source with RXTE twice, during the onset of each active episode. We confirm the pulsations at the 5.16 s period reported earlier (Hurley et al. 1998b, Hurley et al. 1998 e) from SGR 1900+14. Here we report the detection of a secular spindown of the pulse period at an average rate of 1.1*10^{-10} s/s. In view of the strong similarities between SGRs, we attribute the spindown of SGR 1900+14 to magnetic dipole radiation, possibly accelerated by a quiescent flux, as in the case of SGR 1806-20 (Kouveliotou et al. 1998a). This allows an estimate of the pulsar dipolar magnetic field, which is 2-8*10^{14} G. Our results confirm that SGRs are magnetars.

We present a 2 - 10 keV ASCA observation of the field around the soft gamma repeater SGR1900+14. One quiescent X-ray source was detected in this observation, and it was in the SGR error box. In 2 - 10 keV X-rays, its spectrum may be fit by a power law with index -2.2, and its unabsorbed flux is 9.6 x 10^-12 erg cm^-2 s^-1. We also find a clear 5.16 s period. The properties of the three well-studied soft gamma repeaters are remarkably similar to one another, and provide evidence that all of them are associated with young, strongly magnetized neutron stars in supernova remnants.

We present new results based on observations carried out with the Rossi X-ray Timing Explorer during the decay of an outburst of the low-mass X-ray binary (LMXB) and atoll source 4U 1608-52. Our results appear to resolve, at least in 4U 1608-52, one of the long-standing issues about the phenomenology of the kilohertz quasi-periodic oscillations (kHz QPOs), namely, the lack of a unique relation between the frequency of the kHz QPOs and the X-ray flux. We show that despite its complex dependence on the X-ray flux, the frequency of the kHz QPOs is monotonically related to the position of the source in the color-color diagram. Our findings strengthen the idea that, as in the case of Z sources, in the atoll sources the X-ray flux is not a good indicator of $\dot M$, and that the observed changes in the frequency of the kHz QPOs in LMXBs are driven by changes in $\dot M$. These results raise some concern about the recently reported detection of the orbital frequency at the innermost stable orbit in 4U 1820-30.

We discuss the first results of our ongoing program of Target of Opportunity observations of gamma-ray bursts (GRBs) using the SCUBA instrument on the James Clerk Maxwell Telescope. We present the results for GRB 970508, 971214, 980326, 980329, 980519, and 980703. Our most important result to date is the detection of a fading counterpart to GRB 980329 at 850 microns. Although it proved to be difficult to find the infrared counterpart to this burst, the sub-millimeter flux was relatively bright. This indicates that intrinsically the brightness of this counterpart was very similar to GRB 970508. The radio through sub-millimeter spectrum of GRB 980329 is well fit by a power law with index alpha = +0.9. However, we cannot exclude a nu^(1/3) power law attenuated by synchrotron self-absorption. An alpha = +1 VLA-SCUBA power law spectrum is definitely ruled out for GRB 980703, and possibly also for GRB 980519. We cannot rule out that part of the sub-millimeter flux from GRB 980329 comes from a dusty star-forming galaxy at high redshift, such as the ones recently discovered by SCUBA. Any quiescent dust contribution will be much larger at sub-millimeter than at radio wavelengths. Both a high redshift and large dust extinction would help explain the reddening of the counterpart to GRB 980329, and a redshift of z = 5 has been suggested. The large intensity of this burst might then indicate that beaming is important.

We have studied the hard X-ray variability of the soft X-ray transient GRO J0422+32 with BATSE in the 20-100 keV energy band. Our analysis covers 180 days following the first X-ray detection of the source on 1992 August 5, fully covering its primary and secondary X-ray outburst. We computed power density spectra (PDSs) in the 20-50, 50-100, and 20-100 keV energy bands, in the frequency interval 0.002-0.488 Hz. The PDSs of GRO J0422+32 are approximately flat up to a break frequency, and decay as a power law above, with index ~1. The canonical anticorrelation between the break frequency and the power density at the break, observed in Cyg X-1 and other BHCs in the low state, is not observed in the PDSs of GRO J0422+32. We compare our results with those of similar variability studies of Cyg X-1. The relation between the spectral slope and the amplitude of the X-ray variations of GRO J0422+32 is similar to that of Cyg X-1; however, the relation between the hard X-ray flux and the amplitude of its variation is opposite to what has been found in Cyg X-1. Phase lags between the X-ray flux variations of GRO J0422+32 at high and low photon energies, could only be derived during the first 30 days of its outburst. During this period, the variations in the 50-100 keV lag those in the 20-50 keV energy band by an approximately constant phase difference of 0.039(3) rad in the frequency interval 0.02-0.20 Hz.

We have obtained differential time series photometry of the cataclysmic variable GS Pavonis over a timespan of 2 years. These show that this system is deeply eclipsing (~2-3.5 mag) with an orbital period of 3.72 hr. The eclipse depth and out-of-eclipse light levels are correlated. From this correlation we deduce that the disk radius is changing and that the eclipses in the low state are total. The derived distance to GS Pav is 790+/-90 pc, with a height above the galactic plane of 420+/-60 pc. We classify GS Pav as a novalike system.

In recent observations with the Rossi X-Ray Timing Explorer we have detected two simultaneous quasi-periodic oscillation (QPO) peaks in the low mass X-ray binary and atoll source 4U 1735-44. The lower and higher frequency QPOs have frequencies varying between 632 and 729 Hz, and 982 and 1026 Hz, respectively. The fractional rms amplitudes are 3.7 to 8.1% and 5.0 to 5.8%. The frequency separation between the two QPOs changes from 341+-7 Hz to 296+-12 Hz. The inferred mass accretion rate during our observations is relatively low compared to that during the previous observations, where only a single QPO was present. There is weak evidence that the frequency of the QPOs correlates with the mass accretion rate, as observed in other binaries. Five X-ray bursts were recorded with no detectable oscillations with upper limits for the rms fraction of 4% to 13%.

We observed the anomalous X-ray pulsar 4U 0142+61 using the Proportional Counter Array (PCA) aboard the Rossi X-ray Timing Explorer (RXTE) in March 1996. The pulse frequency was measured as f = 0.11510039(3) Hz with an upper limit of df/dt < 4 * 10^(-13) Hz/s upon the short term change in frequency over the 4.6 day span of the observations. A compilation of all historical measurements showed an overall spin-down trend with slope df/dt = (-3.0 +/- 0.1) * 10^(-14) Hz/s. Searches for orbital modulations in pulse arrival times yielded an upper limit of a_x sin i < 0.26 lt-s (99% confidence) for the period range 70 s to 2.5 days. These limits combined with previous optical limits and evolutionary arguments suggest that 4U 0142+61 is probably not a member of a binary system.

We have analyzed data obtained during two observations with the Rossi X-ray Timing Explorer on January 5 and 8, 1997, of the low-mass X-ray binary (LMXB) and atoll source 4U 1636-53. We measure the frequency separation of the two simultaneous kilohertz quasi-periodic oscillations (kHz QPOs) in this source to be 253.7 +/- 4.7 and 246.4 +/- 5.4 Hz, respectively. These values are inconsistent with being equal to 0.5 times the frequency of the 581-Hz oscillations that have been detected previously in 4U 1636-53 during type I bursts. The weighted average discrepancy is 39.5 +/- 3.5 Hz. This result shows that a simple beat-frequency interpretation of the kHz QPOs, in which the frequency of the oscillations detected during type I bursts equals the separation between the two kHz QPOs (or twice that value), is incorrect.

In observations with the Australia Telescope Compact Array we have resolved the radio counterpart of the unusual X-ray binary Cir X-1 into an asymmetric extended structure on arcsecond scales. In order to quantify the asymmetry we have redetermined as accurately as possible both the optical and radio coordinates of the source. The extended emission can be understood as a compact, absorbed core at the location of the X-ray binary, and extended emission up to 2 arcsec to the southeast of the core. The arcsec-scale extended emission aligns with the larger, more symmetric arcmin-scale collimated structures in the surrounding synchrotron nebula. This suggests that the transport of mass and/or energy from the X-ray binary to the synchroton nebula is occurring via the arcsec-scale structures. The ratio of extended flux from the southeast to that from the northwest of the core is at least 2:1. Interpreted as relativistic aberration of an intrinsically symmetric jet from the source, this implies a minimum outflow velocity of 0.1 c. Alternatively, the emission may be intrinsically asymmetric, perhaps as a result of the high space velocity of the system.

The discovery of the peculiar supernova (SN) 1998bw and its possible association with the gamma-ray burst (GRB) 980425$^{1,2,3}$ provide new clues to the understanding of the explosion mechanism of very massive stars and to the origin of some classes of gamma-ray bursts. Its spectra indicate that SN~1998bw is a type Ic supernova$^{3,4}$, but its peak luminosity is unusually high compared with typical type Ic supernovae$^3$. Here we report our findings that the optical spectra and the light curve of SN 1998bw can be well reproduced by an extremely energetic explosion of a massive carbon+oxygen (C+O) star. The kinetic energy is as large as $\sim 2-5 \times 10^{52}$ ergs, more than ten times the previously known energy of supernovae. For this reason, the explosion may be called a `hypernova'. Such a C+O star is the stripped core of a very massive star that has lost its H and He envelopes. The extremely large energy, suggesting the existence of a new mechanism of massive star explosion, can cause a relativistic shock that may be linked to the gamma-ray burst.

We describe a search of archival data from the Burst and Transient Source Experiment (BATSE). The purpose of the search is to find astronomically interesting transients that did not activate the burst detection (or ``trigger'') system onboard the spacecraft. Our search is sensitive to events with peak fluxes (on the 1.024 s time scale) that are lower by a factor of 2 than can be detected with the onboard burst trigger. In a search of 345 days of archival data, we detected 91 events in the 50--300 keV range that resemble classical gamma ray bursts but that did not activate the onboard burst trigger. We also detected 110 low-energy (25--50 keV) events of unknown origins which may include activity from SGR 1806-20 and bursts and flares from X-ray binaries. This paper gives the occurrence times, estimated source directions, durations, peak fluxes, and fluences for the 91 gamma ray burst candidates. The direction and intensity distributions of these bursts imply that the biases inherent in the onboard trigger mechanism have not significantly affected the completeness of the published BATSE gamma ray burst catalogs.

We present new Rossi X-ray Timing Explorer observations of the low-mass X-ray binary 4U 1608-52 during the decay of its 1998 outburst. We detect by a direct FFT method the existence of a second kilohertz quasi-periodic oscillation (kHz QPO) in its power density spectrum, previously only seen by means of the sensitivity-enhancing `shift and add' technique. This result confirms that 4U 1608-52 is a twin kHz QPO source. The frequency separation between these two QPO decreased significantly, from 325.5 +/- 3.4 Hz to 225.3 +/- 12.0 Hz, as the frequency of the lower kHz QPO increased from 470 Hz to 865 Hz, in contradiction with a simple beat-frequency interpretation. This change in the peak separation of the kHz QPOs is closely similar to that previously seen in Sco X-1, but takes place at a ten times lower average luminosity. We discuss this result within the framework of models that have been proposed for kHz QPO. Beat frequency models where the peak separation is identified with the neutron star spin rate, as well as the explanations previously proposed to account for the similar behavior of the QPOs in Sco X-1, are strongly challenged by this result.

The recent discovery of a supernova (SN 1998bw) seemingly associated with GRB~980425 adds a new twist to the decades-old debate over the origin of gamma-ray bursts. To investigate the possibility that some (or all) bursts are associated with supernovae, we performed a systematic search for temporal/angular correlations using catalogs of BATSE and BATSE/{\it Ulysses} burst locations. We find no associations with any of the precise BATSE/{\it Ulysses} locations, which allows us to conclude that the fraction of high-fluence gamma-ray bursts from known supernovae is small ($<$0.2%). For the more numerous weaker bursts, the corresponding limiting fraction of 1.5% is less constraining due to the imprecise locations of these events. This limit ($1.5% \simeq 18$ bursts) allows that a large fraction of the recent supernovae used as a comparison data set (18 supernovae $\simeq 20%$) could have associated gamma-ray bursts. Thus, although we find no significant evidence to support a burst/supernova association, the possibility cannot be excluded for weak bursts.

By comparing positions on a spectral color-color diagram from 10 black hole candidates (BHCs) observed with Ginga (1354-64, 1826-24, 1630-47, LMC X-1, LMC X-3, GS 2000+25, GS 2023+33, GS 1124-68, Cyg X-1, and GX 339-4) with the observed broad-band noise (0.001-64 Hz; BBN) and quasi-periodic oscillation (QPO) variability, we find that the so-called ``Very High State'' is spectrally intermediate to the Soft/High-State and Hard/Low-State. We find a transition point in spectral hardness where the dependence of the BHC QPO centroid frequency (of GS 1124-68 and GX 339-4) on spectral hardness switches from a correlation to an anti-correlation; where the BBN variability switches from High-State to Low-State; and where the spectral hardness of the QPO relative to that of the BBN variability is a maximum. This coincidence of changing behavior in both the QPO and the broad-band variability leads us to hypothesize that the QPO is due to interaction between the physical components which dominate the behaviors of BHCs when they occupy the Hard/Low and Soft/High States. We conclude that these QPO should be observed from BHCs during transition between these two states. Comparison with QPO and BBN behavior observed during the 1996 transition of Cyg X-1 supports this hypothesis. We also report 1-3 Hz QPO observed in GS 2000+25 and Cyg X-1 in the Hard/Low State, and we compare these to the QPO observed in GS 1124-68 and GX 339-4.

The discovery of X-ray, optical and radio afterglows of gamma-ray bursts (GRBs) and the measurements of the distances to some of them have established that these events come from Gpc distances and are the most powerful photon emitters known in the Universe, with peak luminosities up to 10^52 erg/s. We here report the discovery of an optical transient, in the BeppoSAX Wide Field Camera error box of GRB980425, which occurred within about a day of the gamma-ray burst. Its optical light curve, spectrum and location in a spiral arm of the galaxy ESO 184-G82, at a redshift z = 0.0085, show that the transient is a very luminous type Ic supernova, SN1998bw. The peculiar nature of SN1998bw is emphasized by its extraordinary radio properties which require that the radio emitter expand at relativistical speed. Since SN1998bw is very different from all previously observed afterglows of GRBs, our discovery raises the possibility that very different mechanisms may give rise to GRBs, which differ little in their gamma-ray properties.

Pairs of quasi-periodic oscillations (QPOs) at kilohertz frequencies are a common phenomenon in several neutron-star low-mass X-ray binaries. The frequency separation of the QPO peaks in the pair appears to be constant in many sources and directly related to the neutron star spin frequency. However, in Sco X-1 and possibly in 4U 1608-52, the frequency separation of the QPOs decreases with increasing inferred mass accretion rate. We show that the currently available Rossi X-ray Timing Explorer data are consistent with the hypothesis that the frequency separations in all sources vary by amounts similar to the variation in Sco X-1. We discuss the implications for models of the kilohertz QPOs.

We report on R_c and K band observations of the optical counterpart to GRB 970508 with the Fred Lawrence Whipple Observatory (FLWO) 1.2m telescope. Eleven R_c-band observations were obtained on 1997 May 12, and three on May14. The counterpart clearly faded between the two nights. On May 12 there was no evidence for variability (<9%) on 10-70 minute time scales based on 11 R_c-band observations. On May 19 a one hour observation set a limit on the K magnitude of K>18.2. Comparison of these data points with those obtained by other authors shows that the decay of the optical counterpart can be well fitted by a power law of the form f ~ t**-alpha where alpha = 1.22 +/- 0.03 with occasional fluctuations superposed. We note that the decay of the optical counterpart to another burst, GRB 970228, can also be well fitted with a power law with exponent alpha = 1.0/+0.2/-0.5 with occasional fluctuations superposed. These two decay light curves are remarkably similar in form to that predicted by cosmic-fireball models.

We have reconstructed the spectrum of the afterglow of GRB 970508, on May 21.0 UT (12.1 days after the GRB), on the basis of observations spanning the X-ray to radio range. The low-frequency power law index of the spectrum, $\alpha=0.44\pm0.07$ ($F_\nu\propto\nu^\alpha$), is in agreement with the expected value $\alpha=1/3$ for optically thin synchrotron radiation. The 1.4 GHz emission is self-absorbed. We infer constraints on the break frequecies $\nu_{c}$ and $\nu_{m}$ on May 21.0 UT from a spectral transition from $F_{\nu} \sim \nu^{-0.6}$ to $F_{\nu} \sim \nu^{-1.1}$ in the optical passband around 1.4 days. A model of an adiabatically expanding blast wave, emitting synchrotron radiation, in which a significant fraction of the electrons cool rapidly provides a successful and consistent description of the afterglow observations over nine decades in frequency, ranging in time from trigger until several months later.

We report on Westerbork 1.4 GHz radio observations of the radio counterpart to $\gamma$-ray burst GRB~970508, between 0.80 and 138 days after this event. The 1.4 GHz light curve shows a transition from optically thick to thin emission between 39 and 54 days after the event. We derive the slope $p$ of the spectrum of injected electrons ($dN/d\gamma_{e}\propto\gamma_{e}^{-p}$) in two independent ways which yield values very close to $p=2.2$. This is in agreement with a relativistic dynamically near-adiabatic blast wave model whose emission is dominated by synchrotron radiation and in which a significant fraction of the electrons cool fast.

We discuss observations of the quiescent emission from the soft gamma repeater SGR 1806-20 by the Rossi X-ray Timing Explorer. We find that the 2-20 keV RXTE data is consistent with a constant spectral shape during both active bursting periods and periods of relative quiescence, and is best described by a nonthermal (power law) spectral shape. Using archival ASCA data we find that the quiescent spectrum of SGR 1806-20 is well fit over the energy range 1-30 keV by a power law of photon index 2.31 +/- 0.04, with thermal bremsstrahlung and Raymond-Smith models producing much worse fits to the data.

Galactic accretion driven stellar X-ray sources can be divided into groups in different ways. An important division, which covers almost all known X-ray binaries, can be made according to the mass of the donor star: high-mass X-ray binaries and low-mass X-ray binaries. Another distinction (partially overlapping with the previous one) can be made on the basis of the nature of the accreting object: a strongly magnetized neutron star, a neutron star with a weak magnetic field, or a black hole. In this review I describe the properties of these different types of X-ray binaries, and discuss the mass determinations which are the basis for distinguishing accreting neutron stars from black holes.

We report on the results of optical follow-up observations of the counterpart of GRB 970508, starting 7 hours after the event. Multi-color U, B, V, R$_{c}$ and I$_{c}$ band observations were obtained during the first three consecutive nights. The counterpart was monitored regularly in R$_{c}$ until $\sim$ 4 months after the burst. The light curve after the maximum follows a decline that can be fitted with a power law with exponent $\alpha$ = --1.141 $\pm$ 0.014. Deviations from a smooth power law decay are moderate (r.m.s. = 0.15 magnitude). We find no flattening of the light curve at late times. The optical afterglow fluence is a significant fraction, $\sim$ 5%, of the GRB fluence. The optical energy distribution can be well represented by a power law, the slope of which changed at the time of the maximum (the spectrum became redder).

We report on observations of the fading optical counterpart of the gamma-ray burst GRB 970228, made on 4~September~1997 using the STIS CCD on the Hubble Space Telescope. The unresolved counterpart is detected at V=28 +/- 0.25, consistent with a continued power-law decline with exponent -1.14 +/- 0.05. No proper motion is detected, in contradiction of an earlier claim. The counterpart is located within, but near the edge of, a faint extended source with diameter ~0."8 and integrated magnitude 25.7 +/- 0.25. Comparison with WFPC2 data taken one month after the initial burst and NTT data taken on March 13 shows no evidence for fading of the extended emission. After adjusting for the probable Galactic extinction in the direction of GRB 970228 of A_v=0.7, we find that the observed nebula is consistent with the sizes of galaxies of comparable magnitude found in the Hubble Deep Field and other deep HST images, and that only 2% of the sky is covered by galaxies of similar or greater surface brightness. We therefore conclude that the extended source observed about GRB 970228 is most likely a galaxy at moderate redshift, and is almost certainly the host of the gamma-ray burst.

We discovered a single kHz quasi-periodic oscillation (QPO) near 1150 Hz in the Rossi X-ray Timing Explorer X-ray light curve of the low mass X-ray binary and atoll source 4U 1735-44. The rms amplitude of this peak was 2-3%, and the FWHM 6-40 Hz. There are indications that the kHz QPO frequency decreased from 1160 Hz to 1145 Hz when the count rate increased, which would be quite different from what is observed in other atoll sources for which kHz QPOs have been discovered. In the X-ray color-color diagram and hardness-intensity diagram the source traced out the curved branch (the so-called banana branch) which has been found by previous instruments. The kHz QPO was only detected when the source was at the lowest count rates during our observations, i.e. on the lower part of the banana branch. When 4U 1735-44 was at higher count rates, i.e. on the upper part of the banana branch and at higher inferred mass accretion rate with respect to that on the lower part of the banana branch, the QPO was not detected. Besides the kHz QPO we discovered a low frequency QPO with a frequency near 67 Hz, together with a complex broad peaked noise component below 30 Hz. This 67 Hz QPO may be related to the magnetospheric beat-frequency QPO, which is observed on the horizontal branch of Z sources. This idea is supported by the (peaked) noise found in both 4U 1735-44 and Z sources at frequencies just below the QPO frequency.

We present the R_c band light curve of the optical transient (OT) associated with GRB970228, based on re-evaluation of existing photometry. Data obtained until April 1997 suggested a slowing down of the decay of the optical brightness. However, the HST observations in September 1997 show that the light curve of the point source is well represented by a single power law, with a ``dip'', about a week after the burst occured. The exponent of the power law decay is $\alpha$ = --1.10 $\pm$ 0.04. As the point source weakened it also became redder.

GRS 1915+105, one of the two known galactic microquasars, shows an extremely complex variability in the X-ray band, comparable to no other X-ray source in the sky. Making use of RXTE/PCA data, we have analyzed the X-ray spectral distribution throughout the variability. We find that all variations can be attributed to the rapid appearance and disappearance of the inner region of an optically-thick accretion disk. Since the time scale for each event is related to the maximum radius of the disappearing region, the difference in time structure is due to the time distribution of such radii. The observed relation between the extent of the missing inner region of the disk and the duration of an event is in remarkable agreement with the expected radius dependence of the viscous time scale of the radiation-dominated region of an accretion disk.

We use data obtained with the PCA onboard RXTE to analyze the spectral variations of the black hole candidate and superluminal source GRS 1915+105. Our results suggest that despite the complicated structure of the X-ray light curve, all spectral changes observed in this source can be explained as a sudden disappearing of the inner part of the accretion disk around the compact object, followed by a slower re-filling of the emptied region. The duration of an event and the size of the disappearing region fit remarkably well the expected radius dependence of the viscous time scale for the radiation-pressure dominated region of an accretion disk.

Using a new technique to improve the sensitivity to weak Quasi-Periodic Oscillations (QPO) we discovered a new QPO peak at about 1100 Hz in the March 1996 outburst observations of 4U 1608-52, simultaneous with the ~ 600-900 Hz peak previously reported from these data. The frequency separation between the upper and the lower QPO peak varied significantly from $232.7 \pm 11.5$ Hz on March 3, to $293.1 \pm 6.6$ Hz on March 6. This is the first case of a variable kHz peak separation in an atoll source.

Using a new technique to improve the sensitivity to weak Quasi-Periodic Oscillations (QPO) we discovered a new QPO peak at about 1100 Hz in the March 1996 outburst observations of 4U 1608-52, simultaneous with the ~ 600-900 Hz peak previously reported from these data. The frequency separation between the upper and the lower QPO peak varied significantly from $232.7 \pm 11.5$ Hz on March 3, to $293.1 \pm 6.6$ Hz on March 6. This is the first case of a variable kHz peak separation in an atoll source. We discuss to what extent this result could be accommodated in beat-frequency models such as proposed for the kHz QPOs. We measured the rms fractional amplitude of both QPOs as a function of energy, and we found that the relation is steeper for the lower than for the upper frequency peak. This is the first source where such difference between the energy spectrum of the two kHz QPOs could be measured.

We examine what we regard as key observational results on GRB 970228 and GRB 970508 and show that the accumulated evidence strongly suggests that gamma-ray bursts (GRBs) are cosmological fireballs. We further show that the observations suggest that GRBs are not associated with the nuclear activity of active galactic nuclei, and that late-type galaxies are more prolific producers of GRBs. We suggest that GRBs can be used to trace the cosmic history of the star-formation rate. Finally, we show that the GRB locations with respect to the star-forming regions in their host galaxies and the total burst energies can be used to distinguish between different theoretical models for GRBs.

Using the RXTE PCA we discovered dips in the X-ray light curves of the black-hole candidate GRO J1655-40 during outburst. They are short (~minute) and deep (down to ~8% of the out-of-dip intensity). Similar kind of dips were found in 90s measurements of the RXTE ASM during the same outburst. The occurrences of the dips are consistent with the optically determined orbital period, and were found between photometric orbital phases 0.72 and 0.86. This constitutes the first evidence for orbital variations in X-rays for GRO J1655-40. The PCA data indicate that an absorbing medium is responsible for these dips. Using these results we are able constrain the extent of the absorbing medium and the central X-ray source. GRO J1655-40 was in the canonical high state during our PCA observations.

We have used the Fourier cross spectra of Cyg X-1, as obtained with BATSE during a period of almost 2000 days, to estimate the phase (or time) lags between X-ray flux variations in the 20-50 keV and the 50-100 keV bands as a function of Fourier frequency, nu. We find that these lag spectra do not depend on source state, as measured by the fractional rms variations of the X-ray flux. For frequencies well below the Nyquist frequency the variations in the 50-100 keV band lag those in the 20-50 keV band by a time interval proportional to nu^{-0.8}. Binning effects cause the observed hard X-ray time lags to decrease to zero at the Nyquist frequency. Some previous results on time lags were affected by these binning effects. At photon energies above 10 keV the time lags are approximately proportional to the logarithm of the ratio the energies of the passbands used.

During observations with RXTE we discovered two simultaneous kHz QPOs near 500 Hz and 860 Hz in the Z source Cyg X-2. In the X-ray color-color diagram and hardness-intensity diagram (HID) a clear Z track was traced out. A single kHz QPO peak was detected at the left end of the horizontal branch (HB) of the Z track, with a frequency of 731+/-20 Hz and an amplitude of 4.7+0.8-0.6 % rms in the energy band 5.0-60 keV. Further to the right on the HB, at somewhat higher count rates, an additional peak at 532+/-43 Hz was detected with an rms amplitude of 3.0+1.0-0.7 %. When the source moved down the HB, thus when the inferred mass accretion rate increased, the frequency of the higher-frequency QPO increased to 839+/-13 Hz, and its amplitude decreased to 3.5+0.4-0.3 % rms. The higher-frequency QPO was also detected on the upper normal branch (NB) with an rms amplitude of 1.8+0.6-0.4 % and a frequency of 1007+/-15 Hz; its peak width did not show a clear correlation with inferred mass accretion rate. The lower-frequency QPO was most of the time undetectable, with typical upper limits of 2 % rms, no conclusion on how this QPO behaved with mass accretion rate can be drawn. If the peak separation between the QPOs is the neutron star spin frequency (as required in some beat-frequency models) then the neutron star spin period is 2.9+/-0.2 ms (346+/-29 Hz).

In numerous low-mass X-ray binaries kHz quasi-periodic oscillations (kHz QPOs) are detected. We observed the atoll sources GX 9+1 and GX 9+9 with the RXTE satellite. Both sources were on the (upper) banana branch during our observations. No kHz QPOs were detected, with upper limits of 1.3-1.8 %. It is known from other atoll sources (e.g. 4U 1636-53 and 4U 1820-30) that when they are in the upper banana branch the kHz QPOs are not detected. Thus, it remains possible that when GX 9+1 and GX 9+9 are observed longer on the lower banana, or even in the island state, kHz QPOs are detected in these sources.

We report on the results of R band observations of the error box of the gamma-ray burst of August 28, 1997, made between 4 hours and 8 days after this burst occurred. No counterpart was found varying by more than 0.2 magnitudes down to R = 23.8. We discuss the consequences of this non-detection for relativistic blast wave models of gamma-ray bursts, and the possible effect of redshift on the relation between optical absorption and the low-energy cut off in the X-ray afterglow spectrum.

We discuss observations of the soft gamma repeater SGR 1806-20 during the RXTE Target of Opportunity observations made in November 1996. During the ~50 ksec RXTE observation, HEXTE (15-250 keV) detected 17 bursts from the source, with fluxes ranging from 3 x 10^{-9} to 2.2 x 10^{-7} ergs cm^{-2} s^{-1} (20-100 keV). We obtained spectra for the brighter HEXTE by fitting thermal bremsstrahlung and power law functions over the energy range 17 - 200 keV. The best-fit temperatures and photon indices range from 30 - 55 keV and 2.2 - 2.7, respectively. The weighted average temperature of the HEXTE bursts was 41.8 +/- 1.7 keV, which is consistent with previous SGR 1806-20 burst spectra. The persistent emission from SGR 1806-20 was not detected with HEXTE.

Using the RXTE PCA we discovered deep dips in the X-ray light curves of the black-hole candidates GRO J1655-40 and 4U 1630-47 during outburst. Similar kind of dips for GRO J1655-40 were found in 90s measurements of the RXTE ASM during the same outburst. The duration of the dips in both sources is in the order of minutes. The occurrences of the dips observed with the RXTE PCA and ASM in GRO J1655-40 are consistent with the optically determined orbital period, and were found between photometric orbital phases 0.72 and 0.86. This constitutes the first evidence for orbital variations in X-rays for GRO J1655-40. The PCA data indicate that an absorbing medium is responsible for these dips. The X-ray spectra during the dips can be best described by a heavily absorbed component and an unabsorbed component. In the case of GRO J1655-40 we are able constrain the extent of the absorbing medium and the central X-ray source.

We observed the low-mass X-ray binary and Z source GX 17+2 with the Rossi X-ray Timing Explorer during 6-8 Feb 1997, 1-4 Apr 1997, and 26-27 Jul 1997. The X-ray color-color diagram shows a clear Z track. Two simultaneous kHz quasi-periodic oscillations (QPOs) are present in each observation, whose frequencies are well correlated with the position of the source on the Z-track. At the left end of the horizontal branch (HB) only the higher frequency peak is observed, at 645+/-9 Hz, with an rms amplitude of 5.7+/-0.5%, and a FWHM of 183+/-35 Hz. When the source moves down the Z track to the upper normal branch the frequency of the kHz QPO increases to 1087+/-12 Hz, and the rms amplitude and FWHM decrease by a factor 2. Further down the Z the QPO becomes undetectable, with rms upper limits of typically 2.0%. Halfway down the HB a second QPO appears in the power spectra with a frequency of 480+/-23 Hz. The frequency of this QPO also increases when the source moves along the Z track, up to 781+/-11 Hz halfway down the normal branch, while the rms amplitude and FWHM stay approximately constant at 2.5% and 70 Hz. The QPO frequency difference is constant at 293.5+/-7.5 Hz. Simultaneously with the kHz QPOs we detect HB QPOs (HBOs). The simultaneous presence of HBOs and kHz QPOs excludes the magnetospheric beat-frequency model as the explanation for at least one of these two phenomena.

We report on extensive V, R, and i band photometry of the black-hole candidate Nova Sco 1994 (GRO J1655-40) obtained during March 1996, when the source was close to its quiescent pre-outburst optical brightness (V=17.3). From our observations and data taken from the literature we derive a refined ephemeris for inferior conjunction of the secondary star: HJD 2449838.4198(52) + 2.62168(14) x N. We have modeled the V, R, and i band light curves in terms of a Roche lobe filling secondary and an accretion disk around the compact star, the latter described as a flat cylinder with a radial temperature distribution. From the shape of the light curve we constrain the binary inclination and mass of the secondary star to lie in the ranges 63.7-70.7 degrees and 1.60-3.10 Solar masses, respectively. This limits the mass of the black-hole to the range 6.29-7.60 Solar Masses. The mass range we obtained for the secondary star is supported by the results of stellar evolution calculations.

We present the discovery observations for the optical counterpart of the $\gamma$-ray burster GRB 970508 and discuss its light curve in the context of the fireball model. We analyze the HST data for this object, and conclude that any underlying galaxy must be fainter than $R$ = 25.5. We also present a detailed analysis of the HST images of GRB 970228 claimed to show a proper motion of the optical counterpart and conclude that, within the uncertainties, there is no proper motion. We examine several aspects of the neutron-star merger model for $\gamma$-ray bursts. In particular, we use this model to predict the redshift distribution of $\gamma$-ray bursters, and adopting a recent determination of the cosmic star-formation history, we show that the predicted distribution of $\log N - \log P$ relation is consistent with that observed for GRBs.

We have analyzed the spectral variations of the superluminal black-hole X-ray binary GRS 1915+105 by using data obtained with the PCA on the Rossi XTE. We find that, despite the marked differences in the structure and the time scale of variability, all spectral changes can be attributed to the rapid disappearing of the inner region of an accretion disk, followed by a slower re-filling of the emptied region. The time scale for each event is determined by the extent of the missing part of the disk. The observed relation between the duration of an event and the radius of the disappearing region matches remarkably well the expected radius dependence of the viscous time scale for the radiation-pressure dominated region of an accretion disk.

We report on Westerbork 840 MHz, 1.4 and 5 GHz radio observations of the improved IPN-WFC error box of the gamma-ray burst GRB 970111, between 26.4 hours and 120 days after the event onset. In the ~ 13 sq arcmin area defined by the IPN (BATSE and Ulysses) annulus and the published refined BeppoSAX Wide Field Camera error box we detected no steady sources brighter than 0.56 mJy (4 sigma), and no varying radio emission, down to 1.0 mJy (4 sigma). We also report on B, V, R and I band observations of the error box with the 4.2 m William Herschel Telescope at La Palma.

We report three RXTE/PCA observations of the low-mass X-ray binary 4U 0614+09. They show strong (~ 30% rms) band-limited noise with a cut-off frequency varying between 0.7 and 15 Hz in correlation with the X-ray flux. We observe two non-simultaneous 11-15% (rms) kHz peaks near 728 and 629 Hz in the power spectra of two of our observations when the X-ray flux is ~ 10^{-9} erg cm^{-2} s{-1} (2-10 keV), but find no QPO (< 6% rms) when the X-ray flux is half that. We suggest that count rate may not be a good measure for $\dot M$ even in sources as intrinsically weak as 4U 0614+09, and that QPO frequency and noise cutoff frequency track $\dot M$ more closely than count rate. The QPO increases in rms amplitude from ($11 \pm 1.3$) % at 3 keV to ($37 \pm 12$) % at 23 keV; the fractional amplitude of the band-limited noise is energy-independent. This suggests different sites of origin for these two phenomena. The spectrum of the oscillating flux roughly corresponds to a black body with temperature ($1.56 \pm 0.2$) keV and radius ($ 500 \pm 200$) m (other models fit as well), which might indicate the oscillations originate at a small region on the neutron star surface.

Although more than 2,000 astronomical gamma-ray bursts (GRBs) have been detected, and numerous models proposed to explain their occurrence, they have remained enigmatic owing to the lack of an obvious counterpart at other wavelengths. The recent ground-based detection of a transient source in the vicinity of GRB 970228 may therefore have provided a breakthrough. The optical counterpart appears to be embedded in an extended source which, if a galaxy as has been suggested, would lend weight to those models that place GRBs at cosmological distances. Here we report the observations using the Hubble Space Telescope of the transient counterpart and extended source 26 and 39 days after the initial gamma-ray outburst. We find that the counterpart has faded since the initial detection (and continues to fade), but the extended source exhibits no significant change in brightness between the two dates of observations reported here. The size and apparent constancy between the two epochs of HST observations imply that it is extragalactic, but its faintness makes a definitive statement about its nature difficult. Nevertheless, the decay profile of the transient source is consistent with a popular impulsive-fireball model, which assumes a merger between two neutron stars in a distant galaxy.

Using the Rossi X-ray Timing Explorer we have measured 27 microsecond time delays in 830 Hz quasi-periodic oscillations (QPO) between 4-6 keV and 11-17 keV in 4U 1608-52, with high-energy photons lagging low energy photons, and found upper limits to the time delays of 45 microseconds between 2-6.5 keV and 6.5-67 keV in 730 Hz QPO in 4U 0614+091 and 30 microseconds between 8.7-12.4 keV and 12.4-67 keV in 870 Hz QPO in 4U 1636-53. We also find that the cross coherence function between QPO at different energies is > 0.85 with 95% confidence in 4U 1608-52 and 4U 1636-53. If Compton upscattering of low-energy X-rays in a region with an optical depth of a few is responsible for the delays, then the Compton upscattering region is between a few kilometers and a few tens of kilometers in size.

We investigated the properties of the Z pattern of the low-mass X-ray binary GX 17+2 in the X-ray colour-colour diagrams (CDs) and hardness-intensity diagrams (HIDs), and of the power spectra as a function of position in the Z, using all EXOSAT ME data. The Z pattern in the CD does not move to within 2 per cent between observations separated by 1 day to 1.5 years. The power spectra and the characteristics of the Z pattern both support the division of the Z sources into two sub-groups, GX 17+2 being in the group together with Sco X-1 and GX 349+2. It has been noted previously that the maximum frequency of the horizontal branch quasi-periodic oscillations in GX 17+2 is lower than for Cyg X-2, GX 5-1 and GX 340+0. We suggest that an asymmetry in the magnetic field giving rise to polar caps with different emission characteristics may be the origin of this. This may also explain the occurrence of bursts in GX 17+2. We also report the detection of two new bursts in the EXOSAT data of GX 17+2, which we interpret as type I bursts. No regular pulsations are observed during these two bursts nor during the bursts previously detected in GX 17+2 with EXOSAT, with upper limits to the pulsed amplitude of 1.5 per cent for pulsations with a frequency below 512 Hz. Three of the four bursts occurred in the normal branch, while one occurred in the lower part of the flaring branch.

The repetitive X-ray bursts from the accretion-powered pulsar GRO J1744-28 show similarities to the type II X-ray bursts from the Rapid Burster. Several authors (notably Lewin et al.) have suggested that the bursts from GRO J1744-28 are type II bursts (which arise from the sudden release of gravitational potential energy). In this paper, we present another similarity between these sources. Rossi X-ray Timing Explorer observations of GRO J1744-28 show that at least 10 out of 94 bursts are followed by quasi-periodic oscillations (QPO) with frequencies of about 0.4 Hz. The period of the oscillations decreases over their 30--80 s lifetime, and they occur during a spectrally hard ``shoulder'' (or ``plateau'') which follows the burst. In one case the QPO show a modulation envelope which resembles simple beating between two narrow-band oscillations at 0.325 and 0.375 Hz. Using EXOSAT observations, Lubin et al. found QPO with frequencies of 0.039 to 0.056 Hz following 10 out of 95 type II bursts from the Rapid Burster. As in GRO J1744-28, the period of these oscillations decreased over their 100 s lifetime, and they occurred only during spectrally hard ``humps'' in the persistent emission. Even though the QPO frequencies differ by a factor of 10, we believe that this is further evidence that a similar accretion disk instability is responsible for the type II bursts from these two sources.

We have analyzed the X-ray spectra of the microquasar GRS 1915+105, as observed with the PCA on the Rossi XTE, during periods of stable weak emission, outbursts and rapid flaring. We find that the complicated X-ray intensity curve of this source can be described by the rapid removal and replenishment of matter forming the inner part of an optically thick accretion disk, probably caused by a thermal-viscous instability analogous to that operating in dwarf novae, but here driven by the Lightman-Eardley instability. We find that the mass accretion rate in quiescence is 10^-6 solar masses per year. Only a small fraction of the energy liberated by accretion is emitted as radiation. We suggest that most of this energy is advected into the black hole in the high-viscosity state of the outburst cycle.

We report continuous monitoring of Cygnus X-1 in the 1.3 to 200 keV band using ASM/RXTE and BATSE/CGRO for about 200 days from 1996 February 21 to 1996 early September. During this period Cygnus X-1 experienced a hard-to-soft and then a soft-to-hard state transition. The low-energy X-ray (1.3-12 keV) and high-energy X-ray (20-200 keV) fluxes are strongly anti-correlated during this period. During the state transitions flux variations of about a factor of 5 and 15 were seen in the 1.3-3.0 keV and 100-200 keV bands, respectively, while the average 4.8-12 keV flux remains almost unchanged. The net effect of this pivoting is that the total 1.3-200 keV luminosity remained unchanged to within about 15%. The bolometric luminosity in the soft state may be as high as 50-70% above the hard state luminosity, after color corrections for the luminosity below 1.3 keV. The blackbody component flux and temperature increase in the soft state is probably caused by a combination of the optically thick disk mass accretion rate increase and a decrease of the inner disk radius.

We present results of RXTE observations of the low-mass X-ray binary and atoll source 4U 1608$-$52 made over 9 days during the decline of an X-ray intensity outburst in March 1996. A fast-timing analysis shows a strong and narrow quasi periodic oscillation (QPO) peak at frequencies between 850 and 890 Hz on March 3 and 6, and a broad peak around 690 Hz on March 9. Observations on March 12 show no significant signal. On March 3, the X-ray spectrum of the QPO is quite hard; its strength increases steadily from 5 \% at $\sim$2 to $\sim$20 \% at $\sim$12~keV. The QPO frequency varies between 850 and 890 Hz on that day, and the peak widens and its rms decreases with centroid frequency in a way very similar to the well-known horizontal branch oscillations (HBO) in Z-sources. We apply the HBO beat frequency model to atoll sources, and suggest that, whereas the model could produce QPOs at the observed frequencies, the lack of correlation we observe between QPO properties and X-ray count rate is hard to reconcile with this model.

We report the discovery, with NASA's Rossi X-ray Timing Explorer (RXTE), of the first sub-millisecond oscillation found in a celestial X-ray source. The quasi-periodic oscillations (QPO) come from Sco X-1 and have a frequency of approximately 1100 Hz, amplitudes of 0.6-1.2% (rms) and are relatively coherent, with Q up to ~10^2. The frequency of the QPO increases with accretion rate, rising from 1050 to 1130 Hz when the source moves from top to bottom along the normal branch in the X-ray color-color diagram, and shows a strong, approximately linear correlation with the frequency of the well-known 6-20 Hz normal/flaring branch QPO. We also report the discovery of QPO with a frequency near 800 Hz that occurs, simultaneously with the 1100 Hz QPO, in the upper normal branch. We discuss several possible interpretations, one involving a millisecond X-ray pulsar whose pulses we see reflected off accretion-flow inhomogeneities. Finally, we report the discovery of ~45 Hz QPO, most prominent in the middle of the normal branch, which might be magnetospheric beat-frequency QPO.

We report on 40 kiloseconds of ROSAT HRI observations of the globular cluster M92 (NGC 6341) and 20 kiloseconds of observations of the globular cluster M13 (NGC 6205). We find that the low-luminosity (10^{32.5} erg/sec at 7.5 kpc) source previously observed near the core of M92 with the ROSAT PSPC remains unresolved at HRI resolution; we can identify it with M92 with 99 per cent confidence. In M13 we find that the source seen with the ROSAT PSPC lies within the core and is possibly associated with the cluster (96 per cent confidence). We find probabilities of 99.8 per cent (M92) and 98 per cent (M13) for the presence of additional unresolved emission within these globular clusters. We interpret these results in light of current theories regarding the low-luminosity sources.

We present orbital-phase resolved I and K-band spectroscopy of Cygnus X-3. All spectra show emission lines characteristic of Wolf-Rayet stars of the WN subclass. On time scales longer than about one day, the line strengths show large changes, both in flux and in equivalent width. In addition, the line ratios change, corresponding to a variation in spectral subtype of WN6/7 to WN4/5. We confirm the finding that at times when the emission lines are weak, they shift in wavelength as a function of orbital phase, with maximum blueshift coinciding with infrared and X-ray minimum, and maximum redshift half an orbit later. Furthermore, we confirm the prediction -- made on the basis of previous observations -- that at times when the emission lines are strong, no clear wavelength shifts are observed. We describe a simplified, but detailed model for the system, in which the companion of the X-ray source is a Wolf-Rayet star whose wind is at times ionised by the X-ray source, except for the part in the star's shadow. With this model, the observed spectral variations can be reproduced with only a small number of free parameters. We discuss and verify the ramifications of this model, and find that, in general, the observed properties can be understood. We conclude that Cyg X-3 is a Wolf-Rayet/X-ray binary.

We have examined the power spectral behavior of Cygnus X-1 using approximately 1100 days of BATSE data. These data have been searched for correlations between different power spectral features, and for correlations between power spectral features and energy spectral parameters derived from BATSE occultation analysis.

We have examined fifteen days of CGRO/BATSE data, obtained during the first outburst of the black-hole candidate source GRO J1655-40, to search for rapid variability of its X-ray flux. We find no evidence for significant variability of GRO J1655-40 during our observations, with a 2 sigma upper limit to the fractional r.m.s. amplitude in the frequency range 0.03-0.488 Hz of 6.6%. We cannot, on the basis of our observations, determine the source state (low, high, or very-high state) of GRO J1655-40.

We have analyzed approximately 1100 days of Cygnus X-1 hard X-ray data obtained with BATSE to study its rapid variability. We find for the first time correlations between the slope of the spectrum and the hard X-ray intensity, and between the spectral slope and the amplitude of the rapid variations of the hard X-ray flux. We compare our results with expectations from current theories of accretion onto black holes.

We present an X-ray fast-timing and spectral analysis of the type II bursts and the persistent emission (PE) of the Rapid Burster observed with the EXOSAT Medium-Energy Instrument. The Hardness-Intensity and Color-Color Diagrams of the Rapid Burster are somewhat different from those of the majority of other low-mass X-ray binaries, which fall into two distinct groups, ``atoll'' and ``Z'' sources. The strength and frequencies of quasi-periodic oscillations (QPO) in the bursts and the PE, as well as a strong anti-correlation between QPO frequency and burst peak flux, also distinguish the Rapid Burster from atoll and Z sources. The presence, frequency, and strength of QPO in type II bursts are all correlated in some manner to the spectral hardness, as are the strength of QPO and the very low-frequency noise (VLFN) component in the persistent emission.

We present new high-resolution, high signal-to-noise optical spectra of HD 77581, the optical counterpart of the X-ray source Vela X-1, and determine radial velocities from these spectra, as well as from high-resolution IUE spectra and from digitized photographic spectra. The measured velocities show strong deviations from a pure Keplerian radial-velocity curve, which are autocorrelated within one night, but not from one night to another. Since lines of different ions exhibit very similar changes in profile, these deviations most likely reflect large-scale motions of the stellar surface akin to non-radial pulsations. A possible cause could be that the varying tidal force exerted by the neutron star in its eccentric orbit excites high-order pulsation modes in the optical star which interfere constructively for short time intervals. The effect of such velocity excursions on the orbital solution is estimated by means of a Monte-Carlo simulation technique. We investigate sources of systematic error, due to, e.g., the tidal deformation of the star, and find, in particular, evidence for a systematic perturbation of the radial velocity near the time of velocity minimum. This possible distortion severely compromises the accuracy of the radial-velocity amplitude, leading to a 95\% confidence range of 18.0--28.2\,\kms. The corresponding 95\% confidence limits of the masses are given by $\MX=1.9_{-0.5}^{+0.7}\,\Msun$ and $\Mopt=23.5_{-1.5}^{+2.2}\,\Msun$.

We analyze the currently available observations of X-ray binaries in a consistent way, to re-determine the masses of the neutron stars in these systems. In particular, our attention is focussed on a realistic and consistent assessment of observational uncertainties and sources of systematic error. Confidence limits for these new mass estimates are generally less optimistic than previously assumed. The available observations, including data on six radio pulsars, do not firmly constrain the equation of state of neutron star matter. In particular, a firm upper mass limit cannot yet be established. An improvement of the accuracy of optical data holds the key to further progress.