LISM structure — Fragmented superbubble shell?

Small scale structure in local interstellar matter (LISM) is considered. Overall morphology of the local cloud complex is inferred from Ca II absorption lines and observations of H I in white dwarf stars. Clouds with column densities ranging from 2–100 × 10¹⁷ cm^–2 are found within 20 pc of the Sun. Cold (50 K) dense (10⁵ cm^–3) small (5–10 au) clouds could be embedded and currently undetected in the upwind gas. The Sun appears to be embedded in a filament of gas with thickness 0.7 pc, and cross-wise column density 2 × 10¹⁷ cm^–2. The local magnetic field direction is parallel to the filament, suggesting that the physical process causing the filamentation is MHD related. Enhanced abundances of refractory elements and LISM kinematics indicate outflowing gas from the Scorpius-Centaurus Association. The local flow vector and Sco data are consistent with a 4,000,000 year old superbubble shell at –22 km s^–1, which is a shock front passing through preshock gas at –12 km s^–1, and yielding cooled postshock gas at –26 km s^–1in the upwind direction. A preshock magnetic field strength of 1.6 G, and postshock field strength of 5.2 G embedded in the superbubble shell, are consistent with the data.Abbreviations LISM Local ISM - SIC Surrounding Interstellar Cloud - LIC Local Interstellar Cloud     

Spartan 201 coronal spectroscopy during the polar passes of ulysses

Spartan 201 is a shuttle deployed spacecraft that is scheduled to perform ultraviolet spectroscopy and white light polarimetry of the extended solar corona during two 40 hour missions to occur in September 1994 and August 1995. The spectroscopy is done with an ultraviolet coronal spectrometer which measures the intensity and spectral line profile of HI Ly up to heliocentric heights of 3.5 solar radii. It also measures the intensities of the OVI doublet at 1032 and 1037 Å and of Fe XII at 1242 Å. The HI Ly line profile measurements are used to determine the random velocity distribution of coronal protons along the line-of-sight. The absolute HI Ly intensities can be used together with electron densities from the white light coronagraph to estimate electron temperatures from hydrogen ionization balance calculations, and bulk outflow velocities from models of Doppler dimmed resonant scattering. Intensities of minor ion lines are used to determine coronal abundances and outflow velocities of O⁵⁺. Ultraviolet spectroscopy of extended coronal regions from the 11 April 1993 mission of Spartan 201 are discussed.     

Long term observations of Cyg X-3 with Vela 5B

We present long-term (1969–1979) observations of Cygnus X-3, obtained by the Vela 5B satellite. The 3–12 keV light curve has 10 day time resolution. Cyg X-3 is a peculiar high-luminosity X-ray source, radiating from the radio region to hard gamma rays of more than 10¹⁶ eV. It has a 4.8 hour period, probably orbital, which is not resolved by our present analysis. Long term periodicities of 17, 20, and 33–34 days have been reported by several authors, and explained as the effects of apsidal motion, precession, or an eccentric orbit. We do not observe the 17 and 33–34 day variations, and set upper limits significantly lower than the reported amplitude of the 33–34 day variation. There is weak evidence for a 20 day flux variation. The light curve shows high and low states which alternate with a characteristic timescale of 1 year. There is no counterpart, at this time resolution, of the giant radio outburst of 1972 September.     

Recent results on the parameters of the interstellar helium from the ULYSSES/GAS experiment

Velocity and direction of the flow of the interstellar helium and its temperature and density have been determined from the measurements of the ULYSSES/GAS experiment for two different epochs: during the in-ecliptic path of ULYSSES, representing solar maximum conditions, and during the south to the north pole transition (11/94-6/95), close to the solar minimum conditions. Within the improved error bars the values are consistent with results published earlier.The determination of the density n of the interstellar helium at the heliospheric boundary from observations in the inner solar system requires knowledge about the loss processes experienced by the particles on their way to the observer. The simultaneous observation of the helium particles arriving on direct and indirect orbits at the observer provides a tool to directly determine the effects of the loss processes assumed to be predominantly photoionization and — for particles travelling close to the Sun — electron impact ionization by high-energy solar wind electrons.Such observations were obtained with the ULYSSES/GAS instrument in February 1995, before the spaceprobe passed its perihelion. From these measurements values for the loss rates and the interstellar density could be derived. Assuming photoionization to be the only loss process reasonable fits to the observations were obtained for an ionization rate = 1.1 · 10^–7 s^–1 and a density n 1.7 · 10^–2 cm^–3. Including, in addition, electron impact ionization, a photoionization = 0.6 · 10^–7 s^–1 was sufficient to fit both observations, resulting in a density n 1.4 · 10^–2 cm^–3.On leave from Space Research Centre, Warsaw, Poland.     

Latitudinal variation of solar wind velocity

Single station solar wind velocity measurements using the Ooty Radio Telescope (ORT) in India (operating at 327 MHz) are reported for the period August 1992 to August 1993. Interplanetary scintillation (IPS) observations on a large number of compact radio sources covering a latitudinal range of ±80° were used to derive solar wind velocities using the method of fitting a power law model to the observed IPS spectra. The data shows a velocity versus heliographic latitude pattern which is similar to that reported by Rickett and Coles (1991) for the 1981–1982 period. However, the average of the measured equatorial velocities are higher, being about 470 km s^–1 compared to their value of 400 km s^–1. The distribution of electron density variations (N _e ) between 50R and 90R was also determined and it was found that N _e was about 30% less at the poles as compared to the equator.     

The interpretations of ultraviolet observations of comets

Summary Ultraviolet observations of comets from above the Earth's atmosphere have provided excellent new results and improved older ground based observations (OH) by an order of magnitude. Satellites are especially suitable because long integration times and observations during extended time intervals are possible.The existing cometary L observations have confirmed the relatively high overall gas production rates on the order of 10³⁰ molecule s^–1. The results strongly support the concept of an icy conglomerate solid cometary nucleus. Additional observations of hydroxyl and oxygen favor water to be one of the most abundant molecules in comets. The observations are in agreement with the predominent role of water in the evaporation process of the nuclear ices but are not proof in themselves.Water did not outnumber other consitutents by orders of magnitude in comets Bennett and Kohoutek. At least in these comets, carbon-containing molecules were possibly as numerous as water. Determination of the carbon scale length is necessary for a more quantitative statement.A hydrogen velocity of 7–9 km s^–1 was observed in comet Bennett as well as in comet Kohoutek for a variety of heliocentric distances and varying production rates. Determinations of the outflow velocity from L isophotes agree with line profile observations of L and H. Hydroxyl may constitute the main source for the hydrogen atoms with v _H - 8 km s^–1. The decay process, however, leading to this particular velocity is not yet known. Possibly a large portion of the OH radicals do not decay into hydrogen atoms or at least not into slow ones. If the high velocity component of 20 km s^–1 or more comprises a larger amount (up to 50%), most of the quoted hydrogen production rates must be revised upward.The intrinsic cometary brightness is only a very crude indicator of a comet's actual gas production rate as shown by comparison of comets Bennett and TSK. Comets can be successfully used as (extra ecliptic) space probes to measure interplanetary quantities, e.g., the curvature of the extended hydrogen clouds can be used for the absolute determination of the solar emission independent of instrumental calibration. Generally time dependent hydrogen density models must be used for the interpretation. The strength of the ultraviolet L emission favors its measurement as a standard procedure for the observation of comets (possibly together with OH (3090 Å)). These observations provide the most accurate results on the total cometary gas production rate and its variation with heliocentric distance.Dedicated to Professor L. Biermann in recognition of his inspiring guidance.On leave of absence from Max-Planck-Institut für Physik und Astrophysik, Munich.     

The solar grazing incidence (GRIST) and optical (SOT) telescopes joint accommodation

Some of the problems foreseen for the joint accommodation and operation of the Grazing Incidence Solar Telescope (GRIST) under study by ESA to operate in the extreme ultraviolet region (90 < < 1700 Å), and the Solar Optical Telescope (SOT), developed by NASA to operate in the ultraviolet, optical and infrared region (A > 1100 Å) on a Spacelab mission are described.Proceedings of the Conference Solar Physics from Space, held at the Swiss Federal Institute of Technology Zurich (ETHZ), 11–14 November 1980.     

Fast Dust in the Heliosphere

The dynamics of dust particles in the solar system is dominated by solar gravity, by solar radiation pressure, or by electromagnetic interaction of charged dust grains with the interplanetary magnetic field. For micron-sized or bigger dust particles solar gravity leads to speeds of about 30 to 40 km s^–1 at the Earths distance. Smaller particles that are generated close to the Sun and for which radiation pressure is dominant (the ratio of radiation pressure force over gravity F _rad/F _grav is generally termed ) are driven out of the solar system on hyperbolic orbits. Such a flow of -meteoroids has been observed by the Pioneer 8, 9 and Ulysses spaceprobes. Dust particles in interplanetary space are electrically charged to typically +5 V by the photo effect from solar UV radiation. The dust detector on Cassini for the first time measured the dust charge directly. The dynamics of dust particles smaller than about 0.1 m is dominated by the electromagnetic interaction with the ambient magnetic field. Effects of the solar wind magnetic field on interstellar grains passing through the solar system have been observed. Nanometer sized dust stream particles have been found which were accelerated by Jupiters magnetic field to speeds of about 300 km s^–1.     

The X-ray imaging telescopes on Exosat

The instrument configuration and performance characteristics of the X-ray imaging telescopes on EXOSAT are described. The instrument comprises two fully independent Wolter I imaging telescopes. Each telescope can be used in either of two principal modes: (i) an imaging mode with either a position sensitive proportional counter or a channel multiplier array plate in the focal plane, (ii) a spectrometer mode which features a 500 lines/mm and/or a 1000 lines/mm transmission grating as dispersive element.Preliminary results from the calibration of the fully integrated experiment indicate an ultimate angular resolution of 8.5 arc sec full width at half maximum or 17.5 arc sec half-power beam width. The ultimate wavelength resolution in the spectrometer mode ranges from 1Å for wavelengths below 50Å, to 5Å at wavelengths near 300Å.A method for estimating the telescope performance is given which reasonably accounts for the influence of the X-ray source spectrum and the degree of interstellar absorption on the counting statistics.A comparison between EXOSAT and the EINSTEIN telescope in terms of band width/resolution and minimum source detectability shows an enhanced potential for EXOSAT relative to EINSTEIN for sources with T 10⁷K and low column densities (< 4 × 10²⁰cm^–2) and a reduced potential for sources with hard, or heavily cut-off, spectra.     

The complex Hα profile of the hypergiant HR 8752 during 1970–1992

The profiles of H observed during the 1970–1992 period in the binary hypergiant HR 8752 (G0 Ia) are presented. We distinguish five typical H profiles designated as A, B, C, D and E types according to the number of emission and absorption features. The profiles of H are complex and contain several emission and absorption components, with: –130 km/s in emission or absorption, –84 km/s in absorption, –49 km/s in emission and about +6 km/s in emission. All of them are rather stable in radial velocities except of the main absorption component in the P Cygni profile with –84 km/s. The frequency of appearance and the periods of duration of the occurrence of the components is discussed. The duration times range between about 3 to 10 months for various components. The red emission component E₂ is particularly interesting. Possible explanations of its origin are discussed.A long-term acceleration of the absorption component in the P Cygni profile is found; it can be interpreted as monotonous acceleration of the stellar wind.     

Solar physics assessment of future high-resolution observations in the Grazing-Incidence domain

The project for a Grazing Incidence Solar Telescope (GRIST) offers, for the first time, the combinations of high spatial (1) and spectral resolution in the extreme-ultraviolet wavelength range. The 3-dimensional electron density and temperature structure of the transition region and corona will be determined. The dynamics of the structures which make up the corona will be studied. GRIST can be expected to provide definitive improvement in the understanding of the coronal heating problem.Proceedings of the Conference Solar Physics from Space, held at the Swiss Federal Institute of Technology Zurich (ETHZ), 11–14 November 1980.     

The Coronal Helium Abundance Experiment on Spacelab 2

The Coronal Helium Abundance Spacelab Experiment, (CHASE), basically consists of a grazing incidence telescope and spectrometer sensitive over the range 150–1335 Å. Whilst aimed primarily at deriving the solar helium abundance from measurements of coronal resonance scattering, its specification has been extended in order to provide a more general purpose solar XUV facility. The instrument will be flown on the Spacelab 2 Mission, currently scheduled for launch in November 1984.Proceedings of the Conference Solar Physics from Space, held at the Swiss Federal Institute of Technology Zurich (ETHZ), 11–14 November 1980.     

X-Ray spectroscopic investigation of the coronal structure of capella

The binary system Capella (G6 III + F9 III) has been observed on 1979 March 15 and on 1980 March 15–17 with the Objective Grating Spectrometer (OGS) onboard theEinstein Observatory. The spectrum measured with the 1000 l/mm grating covers the range 5–30 Å with a resolution < 1 Å. The spectra show evidence for a bimodal temperature distribution of emission measure in an optically thin plasma with one component 5 million degrees and the other one 10 million degrees. Spectral features can be identified with line emissions from O VIII, Fe XVII, Fe XVIII, Fe XXIV, and Ne X ions. Good spectral fits have been obtained assuming standard cosmic abundances. The data are interpreted in terms of emission from hot static coronal loops rather similar to the magnetic arch structures found on the Sun. It is shown that the conditions required by this model exist on Capella. Mean values of loop parameters are derived for both temperature components.     

Transition region line-shifts in the rebound shock spicule model

Spectral emission lines created in the solar chromosphere — corona transition region show net red-shifts. It has been proposed that this may be the result of the return of spicular material. We simulate a spicule numerically using the rebound shock model and find that the resulting hydrodynamic evolution leads to a perceived up-flow in transition region spectral lines even though the average velocity in the line forming region is directed downward. The explanation for this apparent paradox is found in the correlation between density and velocity in the waves generated by the rebound shock spicule.     

High latitude electrodynamics: Observations from S3-2

The high spatial-temporal resolution of instrumentation on the polar-orbiting S3-2 satellite has allowed a wide variety of measurements of the electrodynamic characteristics of both large- and small-scale structures at high latitudes. Analyses of large scale features observed by S3-2 have shown that: (i) The IMF B _ydependence of polar cap convection, first observed in June 1969 by OGO-6 persists in other seasons. During periods of northward IMF B _zextensive regions of sunward convection may be found in the sunlit polar cap. (ii) In the dawn and dusk MLT sectors >90% of the region 1 currents lie equatorward of the convection reversal line. Potentials across the ionospheric projection of the low-latitude boundary layer are typically a few kV. (iii) The location of extra field-aligned currents, near the dayside cusp and poleward of the region 1 current sheet is dependent on the IMF B _ycomponent. (iv) Simultaneous observations by TRIAD and S3-2 show that sheets of field-aligned current extend uniformly for several hours in MLT, but may have an altitude dependence in the 1000–8000 km range. (v) During magnetic storms ionospheric irregularities occur in regions of poleward density gradients and downward field-aligned currents near the equatorward boundary of diffuse auroral precipitation. In the winter polar cap, density irregularities were also found in regions of highly structured electric fields and soft electron precipitation. (vi) During an intense magnetic storm the auroral zone height-integrated Pederson conductivity was calculated to be in the range 10–30 mho and downcoming energetic electron fluxes accounted for between 50% and 70% of the upward Birkeland currents.Analysis of small-scale structures (latitudinal width < 1°), observed by S3-2, have shown that: (i) Intense meridional electric fields (50–250 mV m^-1) generated by charge separation near the inner edge of the plasma sheet drive intense subauroral convection and are associated with field-aligned currents, on the order of 1–2 A m^-2. (ii) Case studies of discrete arcs in the auroral oval have shown that arcs are associated with pairs of small-scale, field-aligned currents embedded in the large-scale region 1/region 2 field-aligned current sheets. The maximum observed field-aligned current was an upward current of 135 A m^-2, confined to a latitudinal width of 2km and carried by field-aligned accelerated electrons. Return (downward) currents associated with arcs are limited to intensities of 10–15 A m^-2. At this limit the ionospheric plasma becomes marginally stable to the onset of ion-cyclotron turbulence. Two instances of plasma vortices, characteristic of auroral curls, have been observed in the region between the paired current sheets. (iii) Sun-aligned arcs in the polar cap are found in a region of negative electric field divergence, embedded in an irregular electric field pattern. The electrons producing the arcs have a temperature of 200 eV and have been accelerated through potential drops of 1 kV along the magnetic field. Return currents may appear on both sides of polar-cap arcs.     

The infrared space observatory (ISO)

The Infrared Space Observatory (ISO), a fully approved and funded project of the European Space Agency (ESA), is an astronomical satellite, which will operate at wavelengths from 2.5–240 m. ISO will provide astronomers with a unique facility of unprecedented sensitivity for a detailed exploration of the universe ranging from objects in the solar system right out to distant extragalactic sources. The satellite essentially consists of a large cryostat containing at launch over 2000 litres of superfluid helium to maintain the Ritchey-Chrétien telescope, the scientific instruments and the optical baffles at temperatures between 2 K and 8 K. The telescope has a 60-cm diameter primary mirror and is diffraction-limited at a wavelength of 5 m. A pointing accuracy of a few arc seconds is provided by a three-axis-stabilisation system consisting of reaction wheels, gyros and optical sensors. ISO's instrument complement consists of four instruments, namely: an imaging photo-polarimeter (2.5–240 m), a camera (2.5–17 m), a short wavelength spectrometer (3–45 m) and a long wavelength spectrometer (43–196 m). These instruments are being built by international consortia of scientific institutes and have been delivered to ESA for in-orbit operations. ISO will be launched in September 1995 by an Ariane 4 into an elliptical orbit (apogee 70000 km and perigee 1000 km) and will be operational for at least 18 months. In keeping with ISO's role as an observatory, the majority of its observing time is being made available to the general astronomical community via a Call for Observing Proposals.     

Relations between ISM inside and outside the heliosphere

Thanks to remarkable new tools, such as the Goddard High Resolution Spectrograph (GHRS) on board the HST and the EUVE spectrometer on the interstellar side, and Ulysses particle detectors on the heliospheric side, it is possible now to begin to compare abundances and physical properties of the interstellar matter outside the heliosphere (from absorption features in the stellar spectra), and inside the heliosphere (from in situ or remote detection of the interstellar neutrals or their derivatives, the pick-up ions or the Anomalous Cosmic Rays detected by the two Voyager spacecraft).Ground-based and UV spectra of nearby stars show that the Sun is located between two volumes of gas of different heliocentric velocities V and temperatures T (see also Linsky et al, this issue). One of these clouds has the same velocity (V= 25.6 km s^–1 from = 255 and =8) and temperature (6700 K) as the heliospheric helium of interstellar origin probed by Ulysses, and is certainly surrounding our star (and then the Local Interstellar Cloud or LIC). This Identification allows comparisons between interstellar constituents on both sides of the heliospheric interface.Ly-alpha background data (absorption cell and recent HST-GHRS spectra) suggest that the heliospheric neutral H velocity is smaller by 5–6 km s^–1 than the local cloud velocity, and therefore that H is decelerated at its entrance into the heliosphere, in agreement with interaction models between the heliosphere and the ISM which include the coupling with the plasma. This is in favor of a non negligible electron density (at least 0.05 cm³). There are other indications of a rather large ionization of the ambient ISM, such as the ionization equilibrium of interstellar magnesium and of sodium. However the resulting range for the plasma density is still broad.The heliospheric neutral hydrogen number density (0.08–0.16 cm^–3) is now less precisely determined than the helium density (0.013–0.017 cm^–3, see Gloeckler, Witte et al, Mobius, this issue). The comparison between the neutral hydrogen to neutral helium ratios in the ISM (recent EUVE findings) and in the heliosphere, suggests that 15 to 70% of H does not enter the heliosphere. The comparison between the interstellar oxygen relative abundance (with respect to H and He) in the ISM and the heliospheric abundance deduced from pick-up ions is also in favor of some filtration, and thus of a non-negligible ionization.For a significant ISM plasma density, one expects a Hydrogen wall to be present as an intermediate state of the interstellar H around the interface between inside and outside. Since 1993, the two UVS instruments on board Voyager 1 and 2 indeed reveal clearly the existence of an additional Ly-alpha emission, probably due to a combination of light from the compressed H wall, and from a galactic source. On the other hand, the decelerated and heated neutral hydrogen of this H wall has recently been detected in absorption in the spectra of nearby stars (see Linsky, this issue).     

Diffuse cosmic X- and gamma radiation: The isotropic component

SummaryA. Spectral features The ability of the various theories to explain the three main spectral features at 1/4 keV, 60 keV and 1 MeV is summarized in Tables II and III.Clearly, confirmation of the reality of these features, especially the soft X-ray and -ray excesses, is one of the key elements in enabling us to decide between the competing theoretical interpretations.B. Energy requirements None of the proposed interpretations are easily explained in terms of the available energy in cosmic rays (except perhaps the Seyfert galaxy proposal, and this runs into difficulties). It seems that one either has to regard normal galaxies at the present epoch as prolific sources of cosmic rays ( 10⁶⁰ erg/galaxy in protons), as is required by the Brecher-Morrison model, or to argue that at early stages in their evolution far more energy is available than at present. One ends up with much the same energy requirement in this approach.One could conceivably identify such an early phase with the radio galaxy or QSO phenomena: in any event, cosmological evolution plays a major role. Cosmology does ease the energy requirements, but only for the inefficient mechanisms, such as nonthermal bremsstrahlung or ° -production.It seems that one still needs the metagalactic cosmic ray flux to be 10^-2 of the galactic flux in the diffuse inverse Compton models, and 10^-2–10^-4 in the nonthermal bremsstrahlung models.Faced with problems of energetics, one is tempted to turn to the most energetic objects in the Universe, namely Seyfert nuclei and QSO's, to provide the basic energy source, whether directly or indirectly, for the diffuse X-ray background. A direct connection could be more readily investigated when X-ray observations are available of more extra-galactic sources.C. Angular variations Another approach, complementary to that of looking for remote discrete sources, is to seek angular fluctuations, or limits on such fluctuations in the diffuse X-ray background.The best results presently available are those from the X-ray experiment on board OSO 3. Schwartz (1970) reports a limit of I/Ifour percent on small-scale (10°) fluctuations over 10–100 keV over about one-quarter of the sky. If one assumes a astrophysics, namely the origin of cosmic rays, is intimately linked to the origin of the X-ray background.It may well be that no single mechanism suffices to account for the entire spectrum of isotropic X- and -radiation. Nature is sufficiently perverse for there to be a reasonable probability that several different processes are contributing, and considerable ingenuity will be required to ascertain which mechanism, if any, is assigned the dominant role in a given spectral region.This review is based on an invited paper presented at the joint meeting of the A. A. S. Division of High Energy Astrophysics, and the A. P. S. Division of Cosmic Physics, Washington, D. C., 28 April–1 May, 1970     

Cosmological variability of fundamental physical constants

Gamow was one of the pioneers who studied the possible variability of fundamental physical constants. Some versions of modern Grand Unification theories do predict such variability. The paper is concerned with three of the constants: the fine-structure constant , the ratio of the proton massm _p to the electron massm _e, and the ratio of the neutron massm _n tom _e. It is shown on the basis of the quasar spectra analysis, that all the three constants revealed no statistically significant variation over the last 90% of the life time of the Universe. At the 2 significance level, the following upper bounds are obtained for the epoch corresponding to the cosmological redshiftsz2–3: /<1.5×10^–3, m _p/m _p<2×10^–3, and m/m<3×10^–4, where x is a possible deviation of a quantityx from its present value,m=m _p+m _n, and the nucleon masses are in units ofm _e. (According to new observational data which became known most recently, m _p/m _p<2×10^–4) In addition a possible anisotropy of the high-redshift fine splitting over the celestial sphere is checked. Within the relative statistical error 3 < 1% the values of turned out to be the same in various quadrants of the celestial sphere, which corresponds to their equality in causally disconnected areas. However, at the 2 level a tentative anisotropy of estimated / values is found in directions that approximately coincide with the direction of the relic microwave background anisotropy.The revealed constraints serve as criteria for selection of those theoretical models which predict variation of ,m _p orm _n with the cosmological time.     

Tailored analyses of 24 Galactic WN stars

The fundamental properties of 24 Galactic WN stars are determined from analyses of their optical, UV and IR spectra using sophisticated model atmosphere codes (Hillier, 1987, 1990). Terminal velocities, stellar luminosities, temperatures, mass loss rates and abundances of hydrogen, helium, carbon, nitrogen and oxygen are determined. Stellar parameters are derived using diagnostic lines and interstellar reddenings found from fitting theoretical continua to observed energy distributions.Our results confirm that the parameters of WN stars span a large range in temperature (T_*=30–90,000 K), luminosity (log L_*/L=4.8–5.9), mass loss (M=0.9–12×10^–5 M yr^–1) and terminal velocity (v =630–3300 km s^–1). Hydrogen abundances are determined, and found to be low in WNEw and WNEs stars (<15% by mass) and considerable in most WNL stars (1–50%). Metal abundances are also determined with the nitrogen content found to lie in the range N/He=1–5×10^–3 (by number) for all subtypes, and C/N 0.02 in broad agreement with the predictions of Maeder (1991). Enhanced O/N and O/C is found for HD 104994 (WN3p) suggesting a peculiar evolutionary history. Our results suggest that single WNL+abs stars may represent an evolutionary stage immediately after the Of phase. Since some WNE stars exist with non-negligible hydrogen contents (e.g. WR136) evolution may proceed directly from WNL+abs to WNE in some cases, circumventing the luminous blue variable (LBV) or red supergiant (RSG) stage.