Extensive spectral observations of η Carinae over the last cycle, and particularly around the 2003.5 low‐excitation event, have been obtained. The variability of both narrow and broad lines, when combined with data taken from two earlier cycles, reveal a common and well‐defined period. We have combined the cycle lengths derived from the many lines in the optical spectrum with those from broad‐band X‐rays, optical and near‐infrared observations, and obtained a period length of Ppres= 2022.7 ± 1.3 d. Spectroscopic data collected during the last 60 yr yield an average period of Pavg= 2020 ± 4 d, consistent with the present‐day period. The period cannot have changed by more than ΔP/P= 0.0007 since 1948. This confirms the previous claims of a true, stable periodicity, and gives strong support to the binary scenario. We have used the disappearance of the narrow component of He i 6678 to define the epoch of the Cycle 11 minimum, T0= JD 245 2819.8. The next event is predicted to occur on 2009 January 11 (±2 d). The dates for the start of the minimum in other spectral features and broad‐bands are very close to this date, and have well‐determined time‐delays from the He i epoch.

A full description of the 5.5‐yr low excitation events in η Carinae is presented. We show that they are not as simple and brief as previously thought, but a combination of two components. The first, the slow variation component, is revealed by slow changes in the ionization level of circumstellar matter across the whole cycle and is caused by gradual changes in the wind–wind collision shock‐cone orientation, angular opening and gaseous content. The second, the collapse component, is restricted to around the minimum, and is due to a temporary global collapse of the wind–wind collision shock. High‐energy photons (E > 16 eV) from the companion star are strongly shielded, leaving the Weigelt objects at low‐ionization state for more than six months. High‐energy phenomena are sensitive only to the collapse, low energy only to the slow variation and intermediate energies to both components. Simple eclipses and mechanisms effective only near periastron (e.g. shell ejection or accretion on to the secondary star) cannot account for the whole 5.5‐yr cycle. We find anti‐correlated changes in the intensity and the radial velocity of P Cygni absorption profiles in Fe iiλ6455 and He iλ7065 lines, indicating that the former is associated to the primary and the latter to the secondary star. We present a set of light curves representative of the whole spectrum, useful for monitoring the next event (2009 January 11).

We report on the analysis of the Chandra-ACIS data of O, B, and WR stars in the young association Cyg OB2. X-ray spectra of 49 O-stars, 54 B-stars, and 3 WR-stars are analyzed and for the brighter sources, the epoch dependence of the X-ray fluxes is investigated. The O-stars in Cyg OB2 follow a well-defined scaling relation between their X-ray and bolometric luminosities:This relation is in excellent agreement with the one previously derived for the Carina OB1 association. Except for the brightest O-star binaries, there is no general X-ray overluminosity due to colliding winds in O-star binaries. Roughly half of the known B-stars in the surveyed field are detected, but they fail to display a clear relationship between L X and L bol . Out of the three WR stars in Cyg OB2, probably only WR 144 is itself responsible for the observed level of X-ray emission, at a very low log 8.8 0.2The X-ray emission of the other two WR-stars (WR 145 and 146) is most probably due to their O-type companion along with a moderate contribution from a wind-wind interaction zone.

Context. Nearly a dozen star-forming galaxies have been detected in γ-rays by the Fermi observatory in the last decade. A remarkable property of this sample is the quasi-linear relation between the γ-ray luminosity and the star formation rate, which was obtained assuming that the latter is well traced by the infrared luminosity of the galaxies. The non-linearity of this relation has not been fully explained yet. Aims. We aim to determine the biases derived from the use of the infrared luminosity as a proxy for the star formation rate and to shed light on the more fundamental relation between the latter and the γ-ray luminosity. We expect to quantify and explain some trends observed in this relation. Methods. We compiled a near-homogeneous set of distances, ultraviolet, optical, infrared, and γ-ray fluxes from the literature for all known γ-ray emitting, star-forming galaxies. From these data, we computed the infrared and γ-ray luminosities, and star formation rates. We determined the best-fitting relation between the latter two, and we describe the trend using simple, population-orientated models for cosmic-ray transport and cooling. Results. We find that the γ-ray luminosity–star formation rate relation obtained from infrared luminosities is biased to shallower slopes. The actual relation is steeper than previous estimates, having a power-law index of 1.35 ± 0.05, in contrast to 1.23 ± 0.06. Conclusions. The unbiased γ-ray luminosity–star formation rate relation can be explained at high star formation rates by assuming that the cosmic-ray cooling region is kiloparsec-sized and pervaded by mild to fast winds. Combined with previous results about the scaling of wind velocity with star formation rate, our work provides support to advection as the dominant cosmic-ray escape mechanism in galaxies with low star formation rates.