A previous article argued that the antagonism between sight-lines with and without a bump at 2200 \AA\ disappears, and that the observed properties of interstellar extinction can be globally understood, if it is accepted that scattered starlight contaminates the observed spectrum of reddened stars. The present paper develops this new paradigm by providing a better understanding of the characteristics of this scattered light. It further examines the consequences of this revision of interstellar extinction theory for interpreting the 2200 \AA\ bump and the diffuse interstellar bands (DIBs). Two implications are worth noting: (1) the effect of interstellar extinction on cosmic distance estimations needs to be reconsidered; (2) it is unnecessary to invoke hypothetical particles, such as Poly Aromatic Hydrogenated molecules (PAHs), to explain the peculiarities of interstellar extinction. Hydrogen, by far the most abundant constituent of interstellar clouds, should alone account for the three major features of interstellar extinction: the departure of the continuum from linearity, the bump, and the DIBs.

Theory predicts that a plane wave scattered by a thin slab of gas yields, in the forward direction and under specific circumstances, a larger irradiance than would be observed in the absence of the gas. This enhanced Rayleigh scattering depends on the size of the Fresnel zones at the slab location, as seen from the observer's position, and results from the coherence of the scattering. On astronomical scales the exceptional size of Fresnel zones (~1500km) has particular relevance when considering forward-scattered starlight by an interstellar cloud of atomic hydrogen.

The 2200 A bump is a major figure of interstellar extinction. Extinction curves with no bump however exist and are, with no exception, linear from the near-infrared down to 2500 A at least, often over all the visible-UV spectrum. The duality linear versus bump-like extinction curves can be used to re-investigate the relationship between the bump and the continuum of interstellar extinction, and answer questions as why do we observe two different kinds of extinction (linear or with a bump) in interstellar clouds? How are they related? How does the existence of two different extinction laws fits with the requirement that extinction curves depend exclusively on the reddening E(B-V) and on a single additional parameter? What is this free parameter? It will be found that (1) interstellar dust models, which suppose the existence of three different types of particles, each contributing to the extinction in a specific wavelength range, fail to account for the observations; (2) the 2200 A bump is very unlikely to be absorption by some yet unidentified molecule; (3) the true law of interstellar extinction must be linear from the visible to the far-UV, and is the same for all directions, including other galaxies. In extinction curves with a bump the excess of starlight (or the lack of extinction) observed at wavelengths less than lambda=4000 A is due to a large contribution of light scattered by hydrogen on the line of sight. Although counter-intuitive this contribution is predicted by theory. The free parameter of interstellar extinction is related to distances between the observer, the cloud on the line of sight, and the star behind it (the parameter is likely to be the ratio of the distances from the cloud to the star and to the observer). The continuum of the extinction curve or the bump contain no information concerning the chemical composition of interstellar cloud.

This paper corrects and completes a previous study of the shape of the extinction curve in the visible and the value of Rv. A continuous visible/infrared extinction law proportional to 1/{\lambda}^p with p close to 1 ({\pm}0.4) is indistinguishable from a perfectly linear law (p = 1) in the visible within observational precision, but the shape of the curve in the infrared can be substantially modified. Values of p slightly larger than 1 would account for the increase of extinction (compared to the p = 1 law) reported for {\lambda} > 1{\mu}m and deeply affect the value of Rv. In the absence of gray extinction Rv must be 4.04 if p = 1. It becomes 3.14 for p = 1.25, 3.00 for p = 1.30, and 2.76 for p = 1.40. Values of p near 1.3 are also attributed to extinction by atmospheric aerosols, which indicates that both phenomena may be governed by similar particle size distributions. A power extinction law may harmonize visible and infrared data into a single, continuous, and universal, interstellar extinction law.

This article discusses the interstellar extinction curve in the visible and the value of Rv. It is concluded that the visible extinction curve is likely to be linear in the visible, and that indirect estimates of Rv from tentative determinations of Av, infrared, or UV observations are questionable. There is currently no evidence of any variation of Rv with direction. If Rv is close to 3, as it has been inferred from mid-infrared data, starlight in the visible is extinguished by a factor F/F_0=(2.5exp{-2micron/lambda})^{E(B-V)} in the visible. But if the visible wavelength range alone is considered, 4 appears as its most natural and probable value, and F/F_0= exp{-2E(B-V)/lambda}.

A careful examination of Red Rectangle bands which have been considered as diffuse interstellar bands (DIBs) in emission shows that a few are likely to be artifacts in the spectrum. Some others result from atmospheric extinction. Consequences for the Red Rectangle band/DIB associations are examined. I will also comment a striking resemblance between the DIB spectrum and the spectrum of NO2 in the 6150-6250A region. This suggests that some DIBs could be provoked by atmospheric molecules.

This study investigates the relationship between atmospheric extinction and the spectrum of the Red Rectangle nebula on scales of a few to a few tens of Ang. It is found that the fine structure of the nebula's continuum short-ward of 6700A is similar to background spectra, and is thus determined either by atmospheric absorption or by light from HD44179 scattered in the earth atmosphere.

This study establishes that SMC, LMC and Milky Way extinction curves obey the same extinction law which depends on the 2200A bump size and one parameter, and generalizes the Cardelli, Clayton and Mathis (1989) relationship. This suggests that extinction in all three galaxies is of the same nature. The role of linear reddening laws over all the visible/UV wavelength range, particularly important in the SMC but also present in the LMC and in the Milky Way, is also highlighted and discussed.

The hypothesis of an extended red emission (ERE) in diffuse Galactic light (DGL) has been put forward in 1998 by Gordon, Witt and Friedmann who found that scattered starlight was not enough to explain the amount of DGL in the R band, in some high Galactic latitude directions. This paper re-investigates, for high Galactic latitudes, the brightnesses and colours of DGL, integrated star and galaxy light (ISGL), and of the total extrasolar light (ISGL+DGL) measured by Pioneer. Under the traditional assumption that DGL is forward scattering of background starlight by interstellar dust on the line of sight, ISGL and Pioneer have very close colours, as it is found by Gordon, Witt and Friedmann. Pioneer observations at high |b| thus accept an alternative and simple interpretation, with no involvement of ERE in DGL.

This paper examines the variations, because of atmospheric extinction, of broad-band visible spectra, obtained from long-slit spectroscopy, in the vicinity of some stars, nebulae, and one faint galaxy.

This paper presents an analysis of a series of spectra in the Red Rectangle nebula. Only the reddest part of the spectra can safely be attributed to light from the nebula, and indicates Rayleigh scattering by the gas, in conformity with the large angles of scattering involved and the proximity of the star. In the blue, light from HD44179, refracted or scattered in the atmosphere, dominates the spectra. This paper questions the reliability of ground-based observations of extended objects in the blue.

We analyze and fit visible spectra of a red horizon at sunrise. The shape of the spectra consist of a blue continuum followed by a red bump. The reddest spectra are well fitted by the product of a spectrum of extinguished sunlight (Rayleigh extinction + ozone absorption) and 1/lambda^4. The former is essentially the radiation field in the outer atmosphere, at the scattering volume location; the latter corresponds to Rayleigh scattering by the gas. Moving to higher altitudes, a second component, corresponding to the spectrum of a blue sky, must be added. The spectra we have obtained are similar to spectra of red nebulae, suggesting there may be other explanations than an emission process to the red color of some nebulae.

SAOZ (Systeme d'Analyse par Observation Zenitale) is a balloon born experiment which determines the column density of several molecular species from the visible spectrum of sunlight. We will use sequence of spectra collected during a sunset to discuss atmospheric extinction, and the nature of the radiation field in the atmosphere. The radiation field in the atmosphere is, from daylight to sunset, and with a clear sky, dominated by light coming from the direction of the sun. This light is composed of direct sunlight (extinguished by the gas), and of sunlight forward-scattered by aerosols. As the sun sets, aerosol scattering is first perceived towards the UV. It progressively replaces direct sunlight over all of the spectrum. Our analysis permits fixing the main parameters of each component of the radiation field at any time. The fits we find for the extinction of sunlight in the atmosphere must also apply to starlight. Thus, the present work can be used in astronomy to correct ground-based spectral observations for extinction in the atmosphere.

This paper presents new investigations on coherent scattering in the forward direction (orders of magnitude; conservation of energy; dependence of scattered light on geometry and wavelength), and on how scattered light contamination in the spectrum of reddened stars is possibly related to as yet unexplained observations (the diminution of the 2200 A bump when the obscuring material is close to the star, the difference between Hipparcos and photometric distances). This paper then goes on to discuss the fit of the extinction curve, a possible role of extinction by the gas in the far-UV, and the reasons of the inadequacy of the Fitzpatrick and Massa [ApJSS, 72 (1990) 163] fit.

In this paper I review a series of observations which do not agree with the standard interpretation of the extinction curve. The consequence is that light we receive from a reddened star must be contaminated by starlight scattered at very small angular distances from the star. The true extinction curve is a straight line from the near infrared to the far-UV. If so, all interstellar grains models must be questionned. Another conclusion concerns the average properties of interstellar grains which seem much more uniform than previously thought.

I have been too conservative in `A review of the properties of the scattered starlight which contaminates the spectrum of reddened stars' (NewA, 2002, 7/4, 191) concerning the conditions implied by an additional component of scattered light in the spectrum of reddened stars. I will review and simplify these conditions. Implications for the column density of the scatterers will then be discussed. Estimates for coherent Rayleigh scattering by hydrogen are in excellent agreement with the observations.

When a star is observed behind an interstellar cloud of sufficient column density, we do not observe the direct light from the star, which is totally extinguished. Rather, we see only starlight scattered at small angles from the star. I use several papers published in New Astronomy to recount the different steps which permit understanding how, and under which conditions, scattered starlight can be more important than direct starlight in the spectrum of a reddened star. Associated problems -the fit of the extinction curve, the nature of the scatterers (small grains, or, atoms or molecules from the gas), the 2200A bump- are also, briefly, discussed.

The scattered starlight which contaminates the spectrum of a reddened star depends on l as 1/lambda^n (n*4), and must be attributed to small particles. Surprisingly these particules, which have a quasi-isotropic phase function, do not participate to the brightness of the nebulae. The importance of the intensity of the scattered light present in the spectrum of reddened stars can be explained only if the scattering is coherent. The intensity of the scattered light is then proportional to the square of the number of particles, and is concentrated within an angle of 10^{-8}" from the star. The scatterers can be small grains, or, atoms or molecules from the gas. Hydrogen satisfies the conditions imposed to the scatterers. -- La lumiere diffusee qui contamine le spectre des etoiles rougies a une dependance en 1/lambda^n (n>=4), caracteristique de particules petites devant la longueur d'onde. Il est surprenant que ces particules, qui doivent diffuser de facon quasi-isotrope, ne contribuent pas a la brillance des nebuleuses. L'importance de la lumiere diffusee par ces petites particules dans le spectre d'une etoile rougie s'explique seulement si la diffusion est coherente. L'intensite diffusee est alors proportionnelle au carre du nombre de diffuseurs. Vue par l'observateur, elle est concentree dans un angle inferieur a 10{-8}'' de l'etoile. Les diffuseurs peuvent etre des petits grains, ou, des atomes ou molecules du gaz. L'hydrogene satisfait remarquablement aux conditions que les observations imposent aux diffuseurs.

The standard interpretation of the extinction curve does not agree with some observations. The light we receive from a reddened star must be contaminated by starlight scattered at very small angular distances from the star. The true extinction curve is a straight line from the near infrared to the far-UV. If so, all interstellar dust models must be questionned. A second conclusion is that the average properties of interstellar grains may be much more uniform than previously thought. -- L'interpretation standard de la courbe d'extinction de la lumiere par la matiere interstellaire n'est pas en accord avec certaines observations. Cela s'explique seulement si la lumiere des etoiles rougies est contaminee par de la lumiere diffusee a tres faible angle de diffusion. La courbe d'extinction exacte est alors une ligne droite du proche infrarouge jusqu'a l'UV lointain. Cela remet en question tous les modeles de poussiere utilises a l'heure actuelle. Il n'y a egalement pas lieu de supposer de variation des proprietes moyennes des grains d'une region de l'espace a l'autre.

For directions of sufficcient reddening (E(B-V) >~ 0.25), there is a simple relation between the slope of the extinction curve in the far-UV and E(B-V). Regardless of direction, the far-UV extinction curve is proportional to 1/lambda^n exp(-2e(B-V)/lambda) (lambda in micron, n = 4), in accordance with the idea that reddened stars spectra are contaminated by scattered light (Zagury, 2001b). This relation is not compatible with the standard theory of extinction which states that far-UV and visible extinctions are due to different classes of grains. In the standard theory model the two (far-UV and visible) extinctions vary thus independently according to the proportion of each type of grains. In preceding papers I have shown that the standard theory cannot explain UV observations of nebulae, and is contradicted by the UV spectra of stars with very low reddening: for how long shall the standard theory be considered as the interpretation of the extinction curve?

The spectrum of HD46223 was established from the optical to the far UV and normalized by the spectrum of a non reddened star of same spectral type. The resulting spectrum is separated into two components. One is the direct starlight. The second is an additional component of light scattered at small angles. In the optical the spectrum is dominated by direct starlight which decreases exponentially due to the linear extinction. Scattered light begins to be noticeable in the near-UV. The near-UV rise of the scattered light is interrupted in the bump region. The wavelength dependence of the scattered light is established in the far-UV, where scattered light dominates the extinction curve. A 1/lambda^4 dependence is found, proving the presence of grains small compared to UV wavelengths. The mathematic expressions of the different components mentionned above give a good fit to the extinction curve in the direction of the star. On a mathematic standpoint the fit can be completed by a Lorentzien for the 2200Ang. bump region. The physical interpretation of the bump may be more difficult to achieve since the paper shows the possibility that only scattered light is extinguished in the bump region. Consequences for the grain size distribution which is necessary to explain the different aspects of scattering in interstellar clouds, for the value of Rv, and problems raised by this interpretation of the spectrum of HD46223, are considered at the end of this paper.

This paper examines the relationship between spectra of stars of same spectral type with extremely low reddenings. According to the standard theory, the relationship between the spectrum of stars with same spectral type and small, but different reddenings should be different in the optical and in the UV. This difference is not observed: the ratio of the spectra of two stars in directions where the reddening is large enough to be detected and low enough not to give a noticeable 2200Ang. bump is an exponential of 1/lambda from the near-infrared to the far-UV. This result is in conformity with the ideas introduced in preceding papers: the exponential optical extinction extends to the UV, and the spectrum of stars with enough reddening is contaminated by light scattered at close angular distance from the stars. An application will be the determination of the spectrum of a non-reddened star from the spectrum of a star of same spectral type with little reddening.

The comparison, undertaken in preceding papers, of the UV observations of nebulae and of reddened stars reveals contradictory aspects of interstellar extinction. The aim of this paper is to understand the implications hidden behind the apparent contradictions. The questions treated will be: how can small grains with an isotropic phase function make an appreciable contribution in the UV spectrum of a star? Why are small grains not observed in the spectrum of a nebula? How much of starlight can be scattered by large grains in the forward direction?

This paper presents an analysis of the UV spectrum of some nebulae with clearly identified illuminating stars, all observed by the IUE satellite. The data show remarkable properties of the UV spectrum of the nebulae. Each spectrum is the product of the star spectrum and a linear function of 1/lambda. There is no peculiar behaviour in the spectrums at 2200A: no bump created in the spectrum of a nebula and no excess of scattering. When moving away from the star, the surface brightness of a nebula decreases as the inverse of the square of the angular distance to the star. These results can logically be interpreted in terms of scattering of starlight. They imply constant properties of the interstellar grains in the UV and in the directions of space sampled by the nebulae, and probably a strong forward scattering phase function. There is no evidence for any particular type of grain which would specifically extinguish starlight at 2200A. Concerning the UV spectrum of a star, this may imply a revisal of the traditional interpretation of the 2200A bump.

In 1998 Gordon, Witt and Friedmann wrote of detection of ERE in diffuse galactic light at high latitudes. The comparaison of diffuse galactic light and starlight over a few regions of the sky cannot, according to the authors, be fully explained by extinction of starlight. 50% of the red interstellar light needs to come from an interstellar emitting process. Different aspects of Gordon et al.'s data reduction work, and of the model which supports their assertion, will be considered here. Analysis of the 100 micron emission of the regions Gordon et al. have studied will prove that the color of the diffuse galactic light is explained by scattering of the background starlight by dust embedded in nearby cirrus. ERE may not be present in diffuse galactic light, and, more generally, in interstellar space.

In 1986 Witt & Schild studied the optical emission of selected bright nebulae. Comparison of the data to a model led the authors to conclude the nebulae have too high a level of I-band emission to be explained solely by the scattering of starlight. The excess is explained as a broad band I emission which originates in the nebulae. This phenomenum is known as Extended Red Emission (ERE). Given Witt & Schild's data, the model they employed, and the estimate of observational errors, I've reached a different conclusion: Witt & Schild's observations are compatible with classical scattering and therefore cannot be used to prove the existence of ERE.

The interpretation of observations over different wavelength domains, which now exist over a large fraction of the sky, will be used to determine relationships between a nebula and its' illuminating source. The illuminating source of a high latitude nebula must be behind the cloud and e.g. at close angular distance. It is either a point source behind the nebula or the average radiation field created by all the background stars. In the latter case it is possible to estimate the maximum surface brightness a nebula can reach. Illumination by the galactic plane is always negligible in up to date observations. The red color of some nebulae, when they are not HII regions, is more likely to be due to large column densities than to an emission process. Concerning interstellar grain properties, the same data will be used to support the well known property that grains scatter starlight in forward direction.

The spectrum of five stars is analysed and explained by the superimposition of two components. One is the extinction of the direct starlight, with an extinction coefficient A which varies as the inverse of the wavelength across all the UV. This linear dependence of extinction upon the inverse of wavelength prolongs to the UV the well known relation which exist in the optical between extinction and wavelength. The second component is an additional feature, superimposed on the extincted direct starlight. This feature is interpreted as starlight scattered by dust at close angular distance to the star, into the beam of the observation.

We have designed and built a large-throughput dual channel photometer, Diabolo. This photometer is dedicated to the observation of millimetre continuum diffuse sources, and in particular, of the Sunyaev-Zel'dovich effect and of anisotropies of the 3K background. We describe the optical layout and filtering system of the instrument, which uses two bolometric detectors for simultaneous observations in two frequency channels at 1.2 and 2.1 mm. The bolometers are cooled to a working temperature of 0.1 K provided by a compact dilution cryostat. The photometric and angular responses of the instrument are measured in the laboratory. First astronomical light was detected in March 1995 at the focus of the new Millimetre and Infrared Testa Grigia Observatory (MITO) Telescope. The established sensitivity of the system is of 7 mK_RJ s^1/2$. For a typical map of at least 10 beams, with one hour of integration per beam, one can achieve the rms values of y_SZ ~ 7 10^-5 and the 3K background anisotropy Delta T/T ~ 7 10^-5, in winter conditions. We also report on a novel bolometer AC readout circuit which allows for the first time total power measurements on the sky. This technique alleviates (but does not forbid) the use of chopping with a secondary mirror. This technique and the dilution fridge concept will be used in future scan--modulated space instrument like the ESA Planck mission project.