The magnitudes of 201 stars at 2600 Å (1000 Å passband) were derived from two plates of the winter Milky Way obtained with a large field camera. A preliminary investigation of the interstellar reddening allowed us to plot a color-spectral type diagram. Stars of type O seem to be brighter than predicted.

Preliminary results obtained from a sounding rocket launched in December 1968 show a strong ultraviolet excess in the central part of the Andromeda nebula, as compared to the color of the sky background.

The color index of M 31 in this spectral range may be explained by an abnormally high proportion of B stars.

The success of Stecher and Milligan (1962) in obtaining UV photoelectric objective dispersion spectra of stars in 1962 inspired similar work at the Royal Observatory, Edinburgh. For a number of reasons we preferred to build a system using grating dispersion in the convergent beam from a 21-cm diameter paraboloidal mirror (Sudbury, 1969). The same principle of random sky scan from a spinning rocket was employed.

Since 1962 the Space Research Division of the Royal Observatory Edinburgh have been planning and launching a number of rocket experiments aimed at observing the absolute stellar fluxes of early-type stars in the region 1000 Å–3000 Å. Although some earlier attempts were unsuccessful due to rocket failures, good photometric data have been obtained from four successful flights in 1967 and 1968, and it is the purpose of this short report to indicate the nature of the data rather than give a complete scientific analysis.

The astronomical satellite Kosmos 215 was launched in April 1968 to examine ultraviolet and X-ray stellar radiation. During a period of 40 days the spectral region from 1250 Å to 2700 Å was recorded by 7 photon-counters and photomultipliers mounted in the focal planes of similar telescopes with apertures of 70 mm, fields of view 1° and focal ratios about ⅓. The optical axes of all the telescopes were parallel. The satellite was not stabilized but special damping was used to reduce the rotational velocity to 0°.12 sec−1. The optical telescopes scanned the sky and recorded all stars which crossed their fields of view.

Ultraviolet stellar fluxes from 1500 to 3200 Å were from February through August 1965 on OSO II. A 15-cm diameter Gregorian telescope with a stepped grating spectrophotometer provided flux measurements in ten adjacent 180 Å wide band passes.

By comparing fluxes of stars showing B-U color excesses smaller than 0.19 and larger than 0.26, an ultraviolet extinction curve has been derived, which agrees with others which have been published as far as the magnitude effect is concerned, but the shape differs somewhat, the slope at 1/λ = 3.8 μ−1 being somewhat greater than at 1/λ = 5.0 μ−1.

If the stars' fluxes are corrected for reddening, it is found that there is a good agreement between observation and models for stars earlier than B3, but that many of the later type stars have a residual apparently intrinsic reddening.

Ultraviolet solar observations are compared with predictions from a new solar model. From 3600 to 1700 Å there is heavy line blanketing; probably one or more major sources of opacity are missing from the theoretical calculation in this region.

The dividing line between photometry and spectrometry is not always obvious and for the purpose of this review, I will define ultraviolet stellar spectroscopy as observations with sufficient spectral resolution to allow the detection of individual spectral lines and their measurement in terms of wavelength and strength. From an examination of the existing observations this results in a resolution requirement of δλ < 10 Å. Since the best spectral resolution so far obtained is about 1 Å then this places the results to be discussed within the range 1–10 Å. In terms of λ/δλ this corresponds to a range of about 2000–200 and it is important to bear in mind that these represent low resolution spectra. In fact the limit of 200 that I have imposed would rarely be used for spectroscopic studies in ground based observatories where it corresponds, in the notation of the optical astronomer, to a dispersion of about 1000 Å/mm, the resolution limit being set by the photographic plate, typically taken as 20 μ. Hence, even the faintest objects like quasars are usually studied with a dispersion of a few hundred Å mm−1. The fact that such a resolution can be included here is an indication of the exceptionally strong resonance lines which occur in the ultraviolet and which can be detected with such a resolution. On the other hand, the richness of the ultraviolet spectrum is making and will continue to make, demands on improved resolution in order to separate the many features. The best achieved resolution of about 1 Å goes only part-way to solving this problem.

Spectral scans of a number of early-type stars have been made with a three-channel spectrophotometer attached to a 32 cm telescope. The telescope was pointed at the individual program stars with an accuracy of at least 20 sec of arc. The spectral resolution was 10 Å over the range 1150 Å to 4000 Å. The data were recorded in digital form and processed in a computer.

The profile of hydrogen Ly-α, as smeared by the exit slit, was obtained for α Lyr, α CMA, β Ori, and β Tau. For the earlier-type stars observed the Ly-α line of the star is masked by interstellar hydrogen absorption. For ζ Oph and ζ Per the Ly-α absorption is considerably larger than for the other early-type stars. The identification of lines is relatively simple for the light elements but becomes more difficult with the heavier ions since there are many more weaker lines and 10 Å is then insufficient resolution. Extreme P Cygni profiles are present in the earliest stars and are interpreted as mass loss at high velocity.

A full description appears in Astrophys. J.159, Feb. 1970.

A spectrum of ζ Pup extending from 920 Å to 1360 Å with approximately 0.8 Å resolution has been recorded at rocket altitudes. Tentative identification of 38 multiplets below 1100 Å has been made from which it is concluded that all the lines appearing in a model atmosphere (Hickok and Morton, 1968) with Te = 37450K have been detected with the exception of those masked by telluric N2 or strong P Cygni-like profiles. Additional absorption lines indicate a wide range of ionization and excitation entirely consistent with observations in the visible spectral region of similar type stars; they also appear to affect sensibly the energy distribution within the spectrum. From newly detected blue-shifted absorption features produced by the ions Svi (933.4 Å, 944.5 Å), Niv (955.3 Å), Niii (991.0 Å), and Ovi (1033.8 Å) mean radial velocities of 1200, 530, 1800 and 1900 km sec−1 respectively have been derived. It is pointed out that the transition in Niv (955.3 Å) does not originate in the ground state configuration as do the other P Cygni profile transitions, but from an excited level which can decay radiatively to the ground state. It seems likely, therefore, that the profile is generated close to the photosphere, and this, together with previously reported results, constitutes evidence for a positive velocity gradient in the detected portion of the circumstellar envelope. On the basis of available data, ions in the ground state of Nv, Niv (by inference) and Niii are all present in the circumstellar envelope, the abundance of the latter remaining large enough to produce a strong P Cygni profile.

This paper reports on preliminary analysis of spectral scans of early-type stars obtained with the Orbiting Astronomical Observatory. The discussion is confined to the spectra of 50 stars observed with a resolution of approximately 10 Å over the spectral interval from 1050 Å to 2000 Å. Following a qualitative description of the spectra, observed equivalent widths of the Siiv (1400 Å) and Civ (1550 Å) lines are compared with model atmosphere calculations. The results suggest an upward revision of the stellar temperature scale for stars earlier than B3 V.

The ultraviolet spectrum of γ Cas from 1060 to 2120 Å is described.

The UV energy curve of Canopus in the 2400–4000 Å wavelength range has been obtained from spectrograms having a dispersion of 183 Å mm−1. This curve is in good agreement with previous observations and with a model atmosphere. The equivalent width of Mg11 2795 Å, 2802 Å is found to be 22 Å.

Up to the present the chief results of investigations of stellar line spectra in the extreme ultraviolet have been limited to two fairly distinct fields. In the case of the Sun the data have referred mainly to the chromosphere and also to the region of the temperature minimum at the top of the photosphere and to the corona. We may hope that observations will eventually be extended to the chromospheric ultraviolet spectra of other stars. In the case of the early type stars the most spectacular results so far obtained have concerned the evidence for mass loss by several of these stars. I have therefore chosen to spend most of the available time discussing some of the ways of studying from the ground, first the chromospheres of stars other than the Sun and secondly mass loss from stars.

Normal red giants of a given spectral type are shown to be heterogeneous with respect to the following chromospheric features: the Balmer absorption lines, the emission line at H∊, and the double-reversed emission lines at Caii H and K. These chromospheric lines are also shown to be strongly time variable, in at least some red giants, on a time scale of a few months or years. Other chromospheric features that require study lie in the infrared (Hei 10830), the near ultraviolet (Feii emission lines), and the vacuum ultraviolet.

Following the detailed study of four very high luminosity OB stars, a survey has been made for spectroscopic evidence of mass loss in a number of early-type supergiants. A list of spectroscopic criteria is given and the mass loss estimates for 24 stars plotted on the HR diagram. The dependence of the phenomenon on spectral type and luminosity is discussed as well as its significance in terms of stellar evolution.

The observed spectral distribution from a star will, in principle, give information about the parts of the atmosphere from which the radiation comes. Most detailed theoretical studies assume that the atmosphere may be represented by plane parallel layers, but in the case of very strong absorption lines and weak continuous absorption, a case which is encountered in the ultraviolet spectral region of B stars, the radiation received from the centers of the lines may come from an entirely different part of the star than does that from the far wings of the lines or from the continuous spectrum. In the case of strong lines, because stars are three dimensional objects and because a field of motion may occur in the outermost part of the atmosphere, the appearance of the ultraviolet line spectrum may be poorly predicted using the hypothesis of stationary plane parallel layers.

Some numerical examples are presented demonstrating that with the UV resonance lines the opacity in the centre of a line may exceed the continuous opacity by a factor 106–108. A summary is given of the chief factors which should be taken into account in any theory of line formation when the hypothesis of LTE is not valid. A detailed description of the interactions between radiation and matter which can occur is necessary. Some examples of the distribution in energy of the lower energy levels of the ion are presented for typical ions of interest. This distribution and the presence of metastable levels are important factors in determining the significance of non-LTE physics. The paper concludes with remarks indicating that ultraviolet spectra of stars should be obtained with a spectral purity of at east 0.1 Å if progress is to be made in understanding the physical state of stars.

We discuss essentially the cases where molecular fluorescences may be excited by Lyman α and other strong discrete ultraviolet emissions, including lines beyond the Lyman limit. The stars involved are the symbiotic objects and the long period variables. The molecules are H2, N2, O2, NO and CO which have their resonance systems in the ultraviolet.

The rocket spectra taken by Morton et al. (1967) were used as a basis for the investigation. These are in the far ultraviolet region of the Orion supergiants δ, ∊, ζ Orionis. Because of the similarity in spectral type and luminosity of these stars and the great strength of the shifted resonance lines it was possible to derive a mean P Cygni profile (standard error 5 % of the continuum) representing a typical single ultraviolet resonance line for a star of spectral type ~B0, Mv ~ −6.2. This is the ‘observed’ mean profile in Figures 2 and 3.

The rocket-ultraviolet observations of strong Doppler-shifted absorption lines of Siiv, Civ, Nv and other ions in the spectrum of O and B supergiants clearly indicate a high velocity outflow of matter from these stars. The presence of moderate ionisation stages in the stellar wind is conclusive evidence that the flow cannot be due to a high temperature corona as is the case for the solar wind. It is shown that the driving mechanism for the hot-star mass loss is radiation pressure exerted on the gas through absorption in resonance lines occurring at wavelengths near the maximum of the star's continuum flux. In the upper layers of these stars the outward force per gram of matter due to the radiation pressure can greatly exceed the gravitational acceleration making a static atmosphere impossible.

The problem of a steady-state moving reversing layer is formulated and the solution leads to predictions of mass-loss rates as a function of effective temperature and gravity for all hot stars. These results are in substantial agreement with the observations.