Previous studies of ultraviolet spectral classification have been insufficient to establish a comprehensive classification system for ultraviolet spectra of early-type stars because of inadequate spectral resolution. We have initiated a new study of ultraviolet spectral classification of B stars using high-dispersion IUE archival data. High-dispersion SWP spectra of MK standards and other B stars are retrieved from the IUE archives and numerically degraded to a uniform resolution of 0.25 or 0.50 Å. The spectra (in the form of plots or photowrites) are then visually examined with the aim of setting up a two-dimensional classification matrix. We follow the method used to create the MK classification system for visual spectra. The purpose of this work is to examine the applicability of the MK system (and in particular, the set of standard stars) in the ultraviolet, and to establish classification criteria in this spectral region.

Widened photographic image-tube spectra at 50 Å mm−1 dispersion have been obtained for a large number of late-type field dwarfs and giants with well-determined atmospheric parameters and for 35 Hyades dwarfs and 31 Pleiades dwarfs. The spectra have a resolution of 2.5 Å and cover the wavelength range λλ3400–4500 Å. A new quantitative three-dimensional spectral classification system is derived for late-type stars and is calibrated using the sample of field dwarfs and giants with known atmospheric parameters. Diagnostic indices are defined by comparing the counts in the bottoms of two neighboring absorption lines or by comparing the counts in two neighboring pseudocontinuum peaks.

It is found that the Hyades and Pleiades dwarfs and field dwarfs with strong Ca II H and K emission reversals exhibit well-defined anomalies in their diagnostic indices when compared with normal field dwarfs. These results are summarized in Rose (1984). In LaBonte and Rose (1984), it is shown, from spectra of solar magnetic plages, that all spectral and photometric peculiarities of the Hyades discovered to date, including the original “Hyades anomaly” (Crawford 1969; Strömgren et al. 1982), are manifestations of plage activity.

Recent continuum and line data in the far ultraviolet for ι Her (B3IV) and τ Sco (BOV) have been combined with published observations in the near uv and visible spectral regions and interpreted with the aid of the Kurucz (1979) line-blanketed model atmospheres. For ι Her and τ Sco, respectively, interpolated models of (Teff/log g) 17500±500K/3.75±0.15, and 31500±1500K/4.3±0.2 fit the observed continua from 900–6500Å and the profiles of Hγ and Hδ. Adopting these models, solar abundances are suggested for both stars.

Optical-region elemental-abundance analyses were performed for ten sharp-lined main sequence B and A stars. The derived abundances are generally in good agreement with those of the Sun. Multiple high dispersion spectrograms, fully line-blanketed solar composition model atmospheres, optical spectrophotometry, and the most accurate gf values were employed. This study provides initial parameters for studies of these stars in the ultraviolet and a consistent set of values for comparison with abundances of more exotic stars.

Following a previous study of stellar abundances (Pasinetti 1980), a relation has been derived between the DDO cyanogen-strength index and values of [Fe/H] reported in the catalog of Cayrel de Strobel, et al. (1980). From this relation we obtain [Fe/H] = 0.08 for the Hyades. This value is slightly lower than that given by Janes (1979), [Fe/H] = 0.13, and is in better agreement with other recent estimates of the Hyades metallicity (Cayrel de Strobel 1980). This improvement is due to the larger quantity of more precise spectroscopic data used in our correlation.

Spectra of carbon dwarf and giant stars may look very similar, as is demonstrated, e. g., by TT CVn and G77-61. Model calculations show that the bands of C2, CN and CO cannot be used in the range log g = 0……6 for log g determinations, since they are affected by abundance. If, moreover, lines of two ionization stages cannot be observed in any element, the only log g indicator would seem to be the ratio of the MgH band and the Mg I b lines. Simple analytic expressions are derived which demonstrate the relation between spectral appearance, abundances and gravity.

The discrepancy in the derived chemical compositions between B-Al stars and stars of later spectral types can be removed if another Teff-calibration is applied for A and F type stars than the one usually used.

Spectrograms of very high quality have been obtained of Vega and Sirius with the Mount Wilson 100-inch (2.5-m) telescope and coudé spectrograph. Examples of these plates, showing the Be II region in the ultraviolet, are exhibited. The reciprocal dispersion is 0.83 Å/mm (83 nm/m) and the FWHM is about 19 mÅ (1.9 pm). The spectrograms have a trailed width of 3 mm and are on IIa or IIIa emulsions.

We show part of a tracing of π Ceti, an example of a uniform series of high-quality tracings of standard stars covering the wavelength region 3850–4650 Å and derived from 10 Å/mm photographic spectrograms taken with the coudé spectrograph of the Mount Wilson 100-inch reflector.

In order to quantify the errors in the IUE line profiles caused by the order overlap, we have compared line depths in IUE and Copernicus spectra. The excess line depth in IUE spectra suggests that the amount of order overlap is about 32% at 1150Å and decreases to zero at about 1400Å, for spectra extracted with the recent version of IUESISPS (the IUE standard extraction software). The transfer of a spectral feature from one order to the next is below the 5% level.

Based on these results, a correction technique is described.

Lunar occultations provide measurements of stellar angular diameters, leading thus to a determination of their effective temperatures. Basically, the data reduction process relies on two methods: model fitting and deconvolution. “Integrated deconvolution”, presented here, is a derived method of deconvolution and its results can compare with the two methods. The purpose is to determine the variation of the uncovered surface of the source throughout the occultation instead of extracting the brightness profile of the source. The main advantage provided is that computing the signal derivative is no longer necessary. In addition to saving a lot of computing time, the method affords a good estimate of the apparent speed, precise dating, and the angular separation of double stars down to 0.002 arcsec. After being tested against the usual methods, “integrated deconvolution” is now currently used. Some results are presented here.

Fracassini, et al. (1983) suggested a preliminary list of stars suitable for use as standards for methods of determination of stellar apparent diameters and absolute radii. According to their criteria, only stars with at least three methods of determination were considered; moreover, only if the percentage error of the data was <10% were they considered as “standards”. This simple statement is not completely satisfactory, because of some ambiguous cases. For instance, the interferometric method (code 1), Wesselink's method (6H) and Barnes and Evans' (6G) method are not independent.

Several direct and indirect methods for the determination of stellar apparent diameters have been developed in the past; they are summarized in Table 1 by Fracassini, et al. (1981) with a code number and references. So far, no detailed comparison of the methods has been carried out, the main difficulty being the extremely small number of common stars. Nevertheless, from the stars listed in CADARS (Fracassini, et al. 1981), we have obtained some correlations in order to test the reliability of some methods and to define the regions of the HR diagram where they can be applied.

For the calibration of radiation detectors, use is currently made of blackbodies, assuming they satisfy Planck's Law. The first problem considered here is then: how well has this law been checked experimentally? Now, it has been pointed out (Crovini and Galgani 1984) that essentially no new experiments have been made after 1921 (Rubens and Michel 1921), when the data were interpreted as fitting the theoretical law within 1%. But, in fact, this work made use of the value 14300 (in suitable units) of the second radiation constant, while the presently adopted value is 14388. When one inserts this value into the calculations, one finds that, indeed, Planck's Law has been checked only to 3%.

For 12 stars from the list of stars with “standard” angular diameters (Fracassini et al. 1983), effective temperatures, bolometric corrections, radii and luminosities were determined. These stars are included in the stellar spectrophotometric catalog of the Sternberg Astronomical Institute and three of them were used as spectrophotometric standards. A comparison was made of Teff obtained directly using angular diameters from the list of Fracassini et al. (1983) and by means of joint determination of Teff and θ (Blackwell and Shallis 1977). For 7 stars the differences in Teff values don't exceed 1–1.5% and the maximum discrepancies are about 6% for BS 2294, 2943 and 4% for the spectrophotometric standard α Aql (BS 7557). Effective temperature values of α Lyr obtained by these two methods are in the agreement within 0.5%.

The energy distributions of 16 horizontal-branch A-stars and 11 horizontal-branch stars in globular clusters have been measured using the Harvard Scanners at KPNO and CTIO and the Oke multichannel spectrophotometer on the 5-m telescope at Mt. Palomar (Philip and Hayes 1983, Hayes and Philip 1983). Wavelengths between 3400 and 6800 Å were measured and reduced to absolute energy distributions on the system of Hayes and Latham (1975). The internal measuring errors were ± 0.034 mag. per observation for the 15th mag. globular cluster stars and ±0.025 mag. per observation for the 7th to 11th mag. field stars, averaged over all wavelengths. Eleven of the field stars have been observed over nine times each and have low internal measurement errors; these stars plus four globular cluster stars with low internal measurement errors are recommended as secondary standard stars. (See Table I.)

Very precise data on the solar radiation has been obtained by merging: (1) the absolute integrals of the disk-center intensity for 20Å-wide spectral bands, observed in the 1960's by Labs and Neckel (1962, 1963, 1967); (2) the ratios of the mean-to-central intensity following from observations of the center-to-limb variation of (a) all 20Å bands below 6600Å (Neckel and Labs 1984), and (b) the intensities at selected continuum wavelengths (Pierce and Slaughter 1977a, b); and (3) the high-resolution Fourier transform spectra (FTS) obtained by J. Brault at Kitt Peak for the center of the disk and for the irradiance.

A new observing station designed for the absolute calibration of stellar and solar radiation has been built in the last few years at the Pic du Midi Observatory, France, at an elevation of 2860 meters. The stellar observations may begin next summer. The main improvements with respect to the calibration experiments carried out during the last decades are the use of a new type of blackbody source and an optical arrangement which is free of systematic errors in the measurements of the star/blackbody ratio. The aim of the experiment is to establish a set of homogeneous standards (about 10 B and A stars) covering the whole northern sky.

Recent results have shed new light on the status of the calibration of absolute stellar fluxes between 912 and 1200 Å. Observations of hot white dwarfs, subdwarfs and planetary nebula nuclei with the Voyager ultraviolet spectrometers provide evidence that the current calibration agrees very well with extrapolations of IUE energy distributions shortwards of 1200 Å. Voyager observations of main sequence B-stars used as flux calibration sources have revealed that many are variable in brightness in the 912–1200 Å region. We conclude there is no current observational motivation for any revision of the 912 to 1200 Å calibration described by Holberg et al. (1982).

Now that larger telescopes are being used to make photometric measures of very faint stars, fainter standard stars are needed. Most of the stars listed in the standard star list of Crawford and Barnes (1970) are too bright, even for a 1.5 m telescope. Over the past two decades observations have been made of B, A and F-type stars in the magnitude range V = 6 to 13 in the northern and southern hemispheres in a program of four-color photometry of field and blue horizontal-branch stars. An earlier list of secondary standards was published in Philip and Philip (1973). The present paper reports on over twice as many standard stars and the secondary standards are increased by about 20%. The number of observations per star is substantially increased.