The discovery of a flux excess in the far‐ultraviolet (FUV) spectrum of elliptical galaxies was a major surprise in 1969. While it is now clear that this UV excess is caused by an old population of hot helium‐burning stars without large hydrogen‐rich envelopes, rather than young stars, their origin has remained a mystery. Here we show that these stars most likely lost their envelopes because of binary interactions, similar to the hot subdwarf population in our own Galaxy. We have developed an evolutionary population synthesis model for the FUV excess of elliptical galaxies based on the binary model developed by Han et al. for the formation of hot subdwarfs in our Galaxy. Despite its simplicity, it successfully reproduces most of the properties of elliptical galaxies with a UV excess: the range of observed UV excesses, both in (1550 −V) and (2000 −V), and their evolution with redshift. We also present colour–colour diagrams for use as diagnostic tools in the study of elliptical galaxies. The model has major implications for understanding the evolution of the UV excess and of elliptical galaxies in general. In particular, it implies that the UV excess is not a sign of age, as had been postulated previously, and predicts that it should not be strongly dependent on the metallicity of the population, but exists universally from dwarf ellipticals to giant ellipticals.

Abstract. Effective temperatures for 420 stars with spectral types between A0 and K3, and luminosity classes between II and V, selected for a flux calibration of the Infrared Space Observatory, ISO, have been determined using the Infrared Flux Method (IRFM). The determinations are based on narrow and wide band photometric data obtained for this purpose, and take into account previously published narrow-band measures of temperature. Regression coefficients are given for relations between the determined temperatures and the photometric parameters (B2−V 1), (b−y) and (B−V ), corrected for interstellar extinction through use of Hipparcos parallaxes. A correction for the effect of metallicity on the determination of integrated flux is proposed. The importance of a knowledge of metallicity in the representation of derived temperatures for Class V, IV and III stars by empirical functions is discussed and formulae given. An estimate is given for the probable error of each temperature determination.

New VIC CCD photometry, obtained with integration times of 20 s, of the sdOB+degenerate‐dwarf eclipsing binary system AA Dor has provided new complete light curves with an rms scatter about a mean curve of ±0.004 mag. These data are analysed with an improved light2 light‐curve synthesis code to yield more accurate determinations of the radii of both stars, the orbital inclination, and the flux ratio between the two components. These radii are only a little different from the values derived 25 years ago from less complete data, but the uncertainties on these values are improved by a factor of 2. The apparent discrepancy remains between the surface gravity of the sdOB primary star obtained from the light‐curve solution with the published spectroscopic orbit and that obtained from non‐local thermodynamic equilibrium analysis of high‐resolution spectra of the sdOB star. The substantial reflection effect in the system is adequately represented by the light2 code with a bolometric albedo of unity in light curves extending from 0.35 to 2.2 μm. However, there are differences at individual wavelengths in the derived albedo, which may indicate a redistribution of flux from shorter wavelengths into the V and IC passbands.

No abstract

We report the discovery of large-amplitude ( -0.25 mag) pulsations in the bright (V= 12.8) sdB star, PG 1605 + 072. The dominant period is 480 s, but more than 20 periods were present on at least three separate occasions. Frequency analysis of the complete data set yields more than 30 periods. A few of these are harmonics or linear combinations of the strongest modes. Excluding the latter, the periods span a range of almost 400 s, which contrasts with the typical range < 20 s for most other EC 14026 stars.Analysis of multicolour photometry limited any cool companion to being a mainsequence star of type MO or later. Balmer line profile fitting yielded an effective temperature of 32 100 + 1000 K and a log g of 5.25 + 0.10, significantly smaller than in the other stars of the EC 14026 class, and possibly indicative of a more evolved state. The lower gravity is probably responsible for the fact that the pulsation periods and amplitudes are respectively much longer and larger than in other stars of the class. This star is an obvious target for asteroseismological investigation using a multilongitude photometric campaign.