Upload 9307.json with huggingface_hub

9307.json

@@ -0,0 +1,52 @@
+{
+    "9307/astro-ph9307006_arXiv.txt": {
+        "abstract": "\\large  We have examined the effect of the environmental density on  the arm classification of an extensive sample of spiral galaxies included in the Nearby Galaxy Catalog (Tully, 1988a). We have also explored the dependence of the arm class of a galaxy on other factors, such as its blue absolute magnitude and its disk-to-total mass ratio, inferred in the literature either from the gradient  of a good galaxy rotation curve or from a photometric mass decomposition method.  We have found that the arm class is strongly related to the  absolute magnitude in the mid-type spirals (in the sense that  grand design galaxies are, on average, more luminous than  flocculent objects), whilst this relation is considerably weaker in the  early and late types. In general the influence of the local density  on the arm structure appears to be  much weaker than that of the  absolute magnitude. The local density acts essentially in strengthening the arm class--absolute magnitude relation for the  mid types, whereas no environmental density effects are observed in  the early and late types.  Using the most recent estimates of the disk-to-total mass ratio, we do not confirm this ratio to be a significant factor which affects the arm  class; nevertheless, owing to poor statistics and large uncertanties, the issue remains open. Neither a local density effect nor an unambiguous bar effect on the  disk-to-total mass ratio is detectable; the latter finding   may challenge some theoretical viewpoints  on the formation of bar structures.  \\bigskip \\bigskip  {\\it Subject headings:} galaxies: general - galaxies: structure -  galaxies: internal motions - galaxies: clustering ",
+        "introduction": "The degree of symmetry and continuity of spiral arms in galaxies is the basis of the classification system  introduced by  Elmegreen \\& Elmegreen (1982). They assigned galaxies to 12 distinct arm  classes (AC) ranging from AC=1 (fragmented arms with no symmetry) to  AC=12 (two long, sharply defined arms which dominate the appearance of  the galaxy). A few years later the same authors published a catalog  of spiral arm classes of 765 galaxies, in which the original classification  was slightly refined (Elmegreen \\& Elmegreen, 1987).  Several properties have been examined for correlations with AC. Grand design (hereafter G) spirals (i.e. with AC$>$5) are on average larger and more luminous than flocculent (hereafter F) galaxies (i.e. with AC$<$4) (Elmegreen \\& Elmegreen, 1982, 1987).  Late-type spirals are mostly flocculent, irrespective of their  bar-type, whereas in the early-type spiral range the percentage of  G objects increases from  $\\sim$ 40\\% to $\\sim$80\\% as we go from unbarred to barred systems (Elmegreen \\& Elmegreen, 1989). But in  many respects G and F spirals are very similar: they do not  appear to differ in star formation rate --- as  traced  by colours, H$\\alpha$ and ultraviolet fluxes, blue and infrared surface  brighnesses (Elmegreen \\& Elmegreen, 1986) --- in neutral hydrogen content  (Romanishin, 1985), in supernova rate (McCall \\& Schmidt, 1986), in  CO surface brightness (Stark, Elmegreen \\& Chance, 1987), in radio  and soft X-ray emissions (per unit light) (Giuricin, Mardirossian \\& Mezzetti, 1989). Recently, Elmegreen \\& Elmegreen (1990) and Biviano et al.  (1991) found that AC correlates with the outer galaxy rotation curve, in the sense that galaxies with steeper curves tend to have flocculent arm appearance and galaxies with flatter curves tend to be grand design.  This would indicate a difference in the relative disk masses (see,e.g., the  review by Whitmore, 1990), with G galaxies having the largest disk-to-halo mass --- in agreement with the predictions of the standard density wave theory (see, e.g., the textbook by Binney \\& Tremaine, 1987).  So far the assessment of environmental influence on the arm structure is one of the worst known observational aspects of  this topic, since seemingly controversial results have appeared in the literature. It has been claimed that the occurrence of G objects in unbarred spirals is greater in spiral members of pairs and groups than  in relatively isolated systems (Elmegreen \\& Elmegreen, 1982) and that,  within the family of galaxy groups, it is greater in the densest groups (Elmegreen \\& Elmegreen, 1983, 1987). On the other hand, comparing three  samples of interacting galaxies and four samples of galaxy pairs with  three control samples of bright field galaxies, Giuricin et al. (1989)  pointed out that F galaxies are more common in the samples of interacting  or binary systems than in (relatively isolated) field objects. Noting that  several individual interacting systems display disordered optical morphology and spatial distribution of neutral hydrogen (e.g., Sulentic \\& Arp, 1985 and references cited therein), Giuricin et al. (1989) suggested  that interactions can destroy G arm structures, thus favouring the occurrence  of F objects in close galaxy systems. (This is not necessarily in contradiction with previous claims if the authors refer to very strong interactions only.)  On the theoretical side, the role of the environment in arm  morphology is generally believed to be important. As a matter of fact,  since the work of Toomre \\& Toomre (1972), close encounters between galaxies  have been frequently proposed as types of perturbation which can initiate  or maintain global spiral structure (Kormendy \\& Norman, 1979; Toomre, 1981; Sorensen, 1985; Sundelius et al., 1987; Pasha \\& Poliachenko, 1987).  The presently confused observational situation has prompted us to  search for observational evidence of environmental effects on arm structures, using a rigorous assessment of the environmental density; this is lacking in the above-mentioned relevant studies, which are content with a fairly  vague characterization of the environment of the galaxy samples used. In \\S 2  we present the galaxy sample that we have chosen and the parameters that  we have taken  as good indicators of the local density of galaxies  for each galaxy of the sample considered. In \\S 3 we explore whether AC depends on the local density of galaxies, analyzing  also correlations betwen AC and other factors which can influence it, such as the galaxy absolute magnitude and the disk-to-total mass ratio. As a by-product of our study, we also check whether barred and unbarred objects  differ in their disk-to-total mass ratio.  In \\S 4 we summarize our main results. ",
+        "conclusions": "With respect to  earlier studies, our investigation into the factors which influence the AC of a spiral galaxy evidences a considerable  morphological type effect on the dependence of AC on \\mb. The AC is strongly related to \\mb\\ (in the sense that G galaxies tend to be more luminous than F objects) in mid types and, only weakly, in late types; in early types this relation  is very marginal (it is observed only in barred systems). In view of the lack of evidence of enhanced star formation in G galaxies (see, e.g., Elmegreen \\& Elmegreen, 1986; Giuricin et al., 1989), this correlation would suggest that prominent wave modes are more easily generated in bright (large) galaxies. It is not easy to understand why this should hold substantially in mid types only.  For many years the spectacular examples of some well-known G galaxies with nearby companions (like M51 or M81) have been regarded as typical cases of G structures triggered by tidal interactions. Also very recently, a large number of numerical simulations have been devoted to reproducing G structures by means of tidal interactions (see, e.g., the extensive survey of computer simulations by Byrd \\& Howard, 1992). However, our statistical study reveals that, amidst the various mechanisms devised for forming and maintaining spirals --- e.g., i) modes to feedback cycles and amplification at corotation, ii) edges and grooves in the density and/or angular momentum distributions, iii) bar-like or oval potentials, iv) local responses of a galactic disk when forced by a clump, like a giant molecular cloud, v) tidal perturbations (see, e.g., the review by Athanassoula, 1990) --- the last one does not seem to be the dominant physical mechanism at play. As a matter of fact, the influence of the local density on AC is statistically quite weak. Nevertheless, at variance with earlier investigations, we find that the local density acts essentially only in modifying the AC-\\mb\\ relation  for mid types, making it tighter in denser environments, whilst no appreciable density effects are detectable in early and late types. As a consequence, if we select subsamples of bright or faint mid types, we find positive or negative AC--\\rs\\ correlations: in this way, for bright mid types it is indeed true that G spirals are found in denser environments. Furthermore, previous claims about the greater frequency of F galaxies in binary/interacting samples than in field galaxy samples (Giuricin et al., 1989) could be reconciled with the present results if faint galaxies or late types were overabundant in binary/interacting samples (e.g., the interacting sample constructed by Keel et al., 1985, shows an excess of late types).  Using the  most recent estimates of the disk-to-total mass ratio, derived either from the gradients of good, extended rotation curves or from a photometric mass decomposition  method, we do not find any significant influence of this ratio on the AC of spiral galaxies. This finding, which is seemingly at variance with earlier studies (Elmegreen \\& Elmegreen, 1990; Biviano et al., 1991), is affected by poor statistics ($f_1$-values) and large uncertanties ($f_2$-values); a larger sample of good  rotation curves is needed to clarify this issue, which is still to be regarded as an open question.  We detect no local density effect on the disk-to-total mass ratio.  Our finding is consistent with the recent results of the  two-dimensional H$\\alpha$ observations of  Amram et al. (1992a,b). Obtaining rotation curves from the analysis of  two-dimensional velocity fields, these authors found that cluster spirals located in the inner and outer regions of the cluster have  rotation curves of similar shape. They disclaimed  the view (see, e.g., the review by Whitmore, 1990), based mainly on slit spectroscopic  observations, that the rotation curves of spirals near the cluster  center tend to decrease in their outer parts.  No unambiguous bar effect on the disk-to-total mass ratio emerges from our study (a different choice of the data sample gives different results); but, in any case, the results raise some problem for some theoretical scenarios of bar formation (see, e.g., the review by  Bertin, 1991), which would view barred galaxies as containing  more fractional disk mass than unbarred systems.  \\bigskip \\bigskip  The authors thank Massimo Persic, Paolo Salucci and Elena Pian for enlightening discussions and for having kindly provided them with the most recent estimates of disk-to-total mass ratios. The authors are also grateful to Harold G. Corwin Jr. for having provided them with the ninth tape version of the Third Reference Catalogue of Bright Galaxies (RC3). This work was partially supported by the Ministry of University and Scientific and Technological Research (MURST) and by the Italian Research Council (CNR-GNA).  \\newpage \\onecolumn"
+    },
+    "9307/astro-ph9307021_arXiv.txt": {
+        "abstract": "We present new  magnitudes derived from 1.65~\\micron\\ images for 23 galaxies in the Ursa Major cluster.  Magnitudes now exist for all but one spiral meeting our criteria for cluster membership and having \\hi\\ velocity width greater than 187\\kms\\ and inclination greater than 45\\deg.  These spirals fit a Tully-Fisher relation with dispersion in intrinsic magnitudes (after known observational uncertainties and the effect of cluster depth are removed) of 0.36 and a slope of 10.2 $\\pm$ 0.6.  The magnitude dispersion is smaller than found in the Virgo cluster but still significantly larger than claimed by some authors.  We find a hint that the \\tf\\ may turn over at the bright end.  Adding the central surface brightness of the disk as a third parameter flattens the slope of the \\tf\\ and may give a distance estimate with slightly less dispersion, but the significance of the decrease must be tested on an independent sample. ",
+        "introduction": "%  The \\tf\\ (Tully \\& Fisher 1977) is one of the most useful ways to measure distances to spiral galaxies (\\eg, Jacoby \\etal\\ 1992). However, the amount of scatter in the \\tf\\ is still a key issue, one of importance not only for estimating the distance uncertainties but also because the amount of scatter is crucial in estimating the bias in the distances themselves (Teerikorpi 1984, 1987, Bottinelli \\etal\\ 1987).  Several authors have found remarkably small dispersions (\\eg, Freedman 1990, review  by Jacoby \\etal\\ 1992), but Fouqu\\'{e} \\etal\\ (1990) and Peletier and Willner (1991, hereinafter Paper~1) have found intrinsic dispersions among Virgo cluster spirals near 0.5~magnitudes in the blue and 0.4~magnitudes in the infrared.  The natural question is whether the large dispersions are a property of just the Virgo cluster or are inherent in the \\tf\\ itself.  An ideal test case is the Ursa Major cluster.  It lies at nearly the same redshift as Virgo, has plenty of spiral galaxies, and earlier studies (Pierce \\& Tully 1988, hereinafter PT) have indicated a smaller Tully-Fisher dispersion than in Virgo.  This paper presents magnitudes derived from infrared images for a nearly complete sample of Ursa Major spirals.  The primary aim is to investigate the dispersion, but we also examine how best to derive magnitudes and inclinations for Tully-Fisher purposes.  Sample selection is given considerable attention. ",
+        "conclusions": "From a H-band imaging study of a complete sample of galaxies in the Ursa Major cluster, selected on the basis of velocity width we conclude:  \\begin{itemize}  \\item The scatter in the \\tf\\  is least when using circular magnitudes within an optically determined diameter. Infrared isophotal magnitudes give very bad fits owing to the strong relation between velocity width and surface brightness. Using extrapolated total magnitudes the scatter is worse than using circular magnitudes.  It is still possible, however, that total magnitudes measured from deeper images may match or improve upon the circular magnitudes.  \\item The scatter in the \\tf\\  in Ursa Major is larger than can be accounted for by its depth. After taking into account uncertainties in the observations and a cluster depth of 0.17 mag, an intrinsic scatter in the magnitude of an individual galaxy of order 0.36 mag is needed.  \\item The scatter in the Ursa Major cluster is slightly smaller than in Virgo.  This is evidence for subclustering in Virgo, assuming that both clusters are spherical.  \\item The slope of the \\tf\\ in total H-magnitudes is $10.2\\pm 0.6$ (for $H^c_{-0.5}$ magnitudes and $\\Delta V _{20}^c$ velocity widths). Since the stellar $M/L$ in this band is rather insensitive to metallicity, this slope corresponds to $M \\propto V^{4.1 \\pm 0.3}$, in agreement with AHM but not with PT.  \\item Dust absorption might cause scatter of $\\sim$0.1~mag, but we see no direct evidence for it.  \\item The \\tf\\  in Ursa Major appears to turn over at large velocity widths.  \\item  Distance estimates may be slightly improved if the central surface bright\\-ness of the galaxy disk is ad\\-ded as a third parameter.  This means that the \\tf\\ can better be replaced by a magnitude -- velocity width -- surface brightness plane.  The usefulness of this relation should be tested on an independent sample.  \\item There is evidence for a background sub-cluster superposed on the northern part of the Ursa Major cluster at a heliocentric velocity $\\ga$1000\\kms.  If this sub-cluster is real, the intrinsic scatter in the \\tf\\ might be quite small.  \\end{itemize}"
+    },
+    "9307/astro-ph9307007_arXiv.txt": {
+        "abstract": "CBS is a new program for binary system light curve analysis, it generates synthetic light curves for a binary system, accounting for eclipses, tidal distortion, limb darkening, gravity darkening and  reflection; it is also possible to compute the light contribution and eclipses of an accretion disk. The bolometric light curve is generated, as well as curves  for the U,B,V,R,I colour bands. In the following we give a brief description of the first version of the program and show some preliminary results. ",
+        "introduction": "For many years the photometric analysis of  binary sys\\-tems has been performed by the rectification method developed by Russel (1912, 1952); with this method the light curve is analysed in terms of the eclipses of two limb darkened spherical disks by removing off-eclipse effects such as reflection and tidal distortion. The Russel method was improved by different authors, using an analytical approach (Kopal 1959), but the detailed computations of close binary systems light curves, with strong tidal and reflection effects, can be done only with computers, by  numerical calculation. \\par \\smallskip In the seventies, when powerful com\\-puters became available to astronomers, many light synthesis programs where written by different authors, such as Wilson and Devinney (1971), Wood (1971), Lucy (1968), Linnel (1984) and others. The most popular of these is the Wilson-Devinney pro\\-gram, which represents the stars, deformed by ti\\-dal and rotational forces, by means of equipotential Roche surfaces, assuming central condensation, synchronous rotation and circular orbits. The gravity darkening effect is considered. Reflection is treated in an approximate way: the geometry is simplified by con\\-sider\\-ing the irradiating surface as a point source, and a term is introduced in the albedo coefficient to account for its shape. For the eclipse computation the surfaces of the two stars are represented by a number of surface elements ( usually about thousand ); for each phase the star closer to the observer is analysed first and an analytical approximation for  the boundary of the visible part of the star is computed from the boundary surface elements. This is used to exclude the eclipsed part of the second star when summing the light of all the visible surface elements. The differential correction method is used to obtain the best parameters for the light curve. \\par \\smallskip The Wilson Devinney program gives a satisfactory solution for the light curves of most close binary systems; but most of the modern interest in close binary systems is focused on accretion disks fenomenology and stars with a collapsed companion. A Wilson type approach can't handle accretion disks and complex geometrical configurations; in fact only few attempts have been done up to now in representing the disk contribution to the light curve (Wilson, Caldwell 1978; Horne 1985). \\par \\smallskip In order to consider the contribution of hot spots and accretion disks we have written a new light synthesis program, which uses a different approach, suited to handle complex geometrical configurations. We are still developing the program; the present version isn't optimized, doesn't treat overcontact bi\\-naries and hot spots, doesn't contain a minimum finding  algorithm to look for the best parameters; all these features will be included in a future version. ",
+        "conclusions": ""
+    },
+    "9307/gr-qc9307010_arXiv.txt": {
+        "abstract": "We compare the predictions of linearized theory for the radiation produced in the collapse of a spherically symmetric scalar field with a full numerical integration of the Einstein equations. We find power-law tails and quasinormal ringing remarkably similar to predictions of linearized theory even in cases where nonlinearities are crucial. We also show that power-law tails develop even when the collapsing scalar field fails to produce a black hole. ",
+        "introduction": "Linearized perturbation theory has been the main analytical -- and until comparatively recently, numerical -- tool for analyzing nonspherical gravitational collapse.  The complexity of the problem has usually made this approach necessary, and it has been assumed until recently that the approach was sufficient.  Recently, however, G\\'omez and Winicour \\cite{GoWi} have focused attention on the extent to which these results are even qualitatively representative of the late stages of collapse.  In the picture given by linear perturbation theory of the late stages of collapse, there are two features which are noteworthy. One is the development of ``quasinormal (QN) oscillations,'' damped oscillations at complex frequencies characteristic of the mass of the black hole background.  The second feature is the decay in time $t$ of perturbations as $1/t^n$, the ``power-law tails.'' There are good reasons to examine more carefully whether these features also appear in the fully nonlinear case. The arguments for the QN oscillations and for the tails are somewhat different, and should be considered separately.  According to linearized theory the QN frequencies are fixed complex numbers multiplied by the inverse of the mass of the black hole background.  (We use here and throughout units in which $c=G=1$.) It seems reasonable that the phenomenon of QN ringing will be a feature of nonlinear collapse. One argument is that some numerical investigations of solutions of the fully nonlinear equations have shown QN ringing to be common \\cite{AbBeHoSeSm}. Secondly, the idea of QN ringing seems ``robust.'' It is a natural frequency associated with a radiative boundary condition, and can occur in many different radiative systems.  If QN ringing is found, to what black hole mass does it correspond? The QN oscillations themselves carry energy and may change the meaning of the mass. A reasonable guess, at least, can be made that the QN frequency evolves somewhat during the collapse.    The situation for the power-law tails is quite different. These tails are not familiar or common phenomena. The explanation of their existence can be given in two very different ways: (i) They can be viewed as the result of the scattering of gravitational waves off the ``effective curvature potential'' of the black hole spacetime \\cite{Pr}, or (ii) they can be associated with the branch cut in the Green function for the wave propagation problem \\cite{Le}. Both arguments leave open the possibility that the tails are idiosyncrasies of the linear approximation.  If QN oscillations or tails are missing from a fully nonlinear collapse, or if there is any significant new qualitative feature, the result might be to undermine confidence in the picture of collapse given to us by the analysis of linear perturbations of black hole backgrounds.  G\\'omez and Winicour \\cite{GoWi} have addressed this question with numerical studies of the collapse of a spherically symmetric scalar field due to its own gravitational pull. Since the spherically symmetric problem involves only a 1+1 hyperbolic system, it is enormously easier to solve numerically than the problem of nonspherical collapse.  What is more, the problem is a wonderful testing ground for comparing nonlinear results and the predictions of linearized theory. In linearized perturbation theory the evolution of a scalar field is governed by essentially the same mathematics that governs the dynamics of nonspherical perturbations. In particular, perturbation theory makes very specific predictions about QN ringing and tails for perturbative spherically symmetric scalar fields.  It is therefore of great interest that in their initial numerical studies of scalar field collapse, G\\'omez and Winicour have seen neither QN ringing nor tails.  In this paper we will study the fully nonlinear evolution of a scalar field minimally coupled to general relativity. We consider first the evolution of a spherically collapsing scalar field; in addition, we consider nonspherical perturbations of this spacetime. We establish that the QN frequencies and the power-law tails of the numerical simulations are in remarkable agreement with the predictions of linearized theory when a black hole develops. If a black hole does not develop, and all energy eventually radiates away to infinity, we find that power-law tails still form. The existence of tails, but not QN oscillations, when holes do not form agrees with the analysis presented in the companion paper, hereafter referred to as Paper I.  The organization of this paper is as follows. In section II we describe the coordinate system and the version of the field equations we use. In section III we describe our discretization and discuss the numerical error.  In section IV we study the collapse of a scalar field for various initial configurations. In section V we study the evolution of multipole moments of test fields on the  collapsing background. In all cases comparisons with the linearized results are made. We end in section VI with a brief summary and with conclusions. ",
+        "conclusions": "In Paper I we revisited the argument \\cite{Pr} for the existence of power-law tails of perturbations of Schwarzschild spacetime, in order to point out that it predicts these tails not only at timelike infinity, but also at null infinity and on the horizon.  Reformulating the argument yielded an important spin-off prediction: Power-law tails should form in general not only in perturbations of black holes. The analysis suggested that the tails should form for massless perturbations of any approximately spherically symmetric spacetime whose metric is approximately Schwarzschild, at least on an outgoing null cone of finite thickness (i.e., finite range of advanced time).  The central idea, as elaborated in Paper I, is quite simple. Radiation going out along a thick null cone (the ``initial burst'' of radiation) will be scattered back in by the spacetime curvature at arbitrarily large radius.  The backscattered radiation then reaches a small radius again at arbitrarily large time, attenuated by a power of Bondi time. This backscattered radiation then evolves tails.  The leading effect in this evolution is independent of spacetime curvature. The exponent of the tails is therefore the same if at late times there is, at small $r$, a star, empty space, or a black hole.  In our numerical work described here we first set out to verify the existence of the tails, as well as of QN ringing, in a collapse situation. We chose the model of a spherically symmetric self-gravitating massless minimally coupled scalar field, because tails and QN oscillations seemed to be absent in the results of a previous investigation\\cite{GoWi}. The explanation of that absence seems to be simply that one has to go to very late times to see the late-time features. We did find tails and QN ringing precisely as expected. Fortunately these features arise about an order of magnitude in time before our code is stopped by a diverging redshift.  Extending the scope of previous work, we paid attention to the late-time behavior of solutions which are subcritical, just below or well below the margin of black hole formation, and we again found power-law tails, but not QN ringing. For what can be considered ``generic\" data (see Paper I), we found that QN ringing disappeared abruptly at the margin of collapse, while the appearance of power-law tails was continuous and unremarkable across  the transition from subcritical to supercritical models. This last observation in itself is a very strong corroboration of the simple picture of tail formation set out above. Further evidence can be found in the dependence of the tail amplitude on the amplitude of the initial data, as shown in Fig.~14. The Bondi mass of the spacetime scales as the square of the amplitude of the initial data and we find that the tail amplitude scales as the cube. This suggests that the tails are scattered off the spacetime curvature only once, picking up a single factor of mass.     In a second extension we considered non-spherically symmetric test fields on the dynamical backgrounds generated by collapsing scalar fields. We found that the exponent of the tails varies with multipole index roughly as for fixed backgrounds. Furthermore we used the fact that QN ringing is less damped in the higher modes to make precise measurements of QN frequencies. For models in which the mass of the background was reasonably constant we found excellent agreement with the predictions on a fixed Schwarzschild background.  For models with significant mass loss we found a shift in QN frequency corresponding to the changing mass of the spacetime.    In summary, we found that the predictions for power-law tails of perturbations of Schwarzschild spacetime \\cite{Pr} hold to reasonable approximation, even quantitatively, in a variety of situations to which the predictions might seem initially not to apply.    In the interest of brevity and timeliness, the results reported here do not exhaust the interesting questions that might be asked.  It will be particularly interesting to explore in more detail the behavior of power-law tails and QN ringing on the critical boundary between collapsing and noncollapsing initial data. It should be said here that our code is probably capable only of much coarser accuracy than that of Choptuik \\cite{Chop}, and will have to be modified for this purpose. On the other hand, our code was adequate for verifying, for our two families of initial data, one of Choptuik's crucial results: that for marginal black hole formation the mass of the hole depends in a universal way on the parameter of the family (in our case the amplitude). See Fig.~15.  \\thanks  We wish to thank Jeffrey Winicour and Roberto G\\'omez for discussions and for making available their results. This work was supported in part by grant NSF PHY92-07225 and by research funds of the University of Utah. J.P. acknowledges hospitality and support from the Institute for Theoretical Physics at UCSB and the National Science Foundation grant PHY89-04035."
+    },
+    "9307/astro-ph9307018_arXiv.txt": {
+        "abstract": "One still cannot conclusively assert that the universe underwent a neutral phase, despite the new COBE FIRAS limit $y <2.5\\times 10^{-5}$ on Compton $y$-distortions of the cosmic microwave background. Although scenarios where the very early ($z\\sim 1000$) ionization is thermal (caused by IGM temperatures exceeding $10^4$K) are clearly ruled out, there is a significant loophole for cosmologies with typical CDM parameters if the dominant ionization mechanism is photoionization. If the ionizing radiation has a typical quasar spectrum, then the $y$-constraint implies roughly $h^{4/3}\\Ob \\Omega_0^{-0.28}<0.06$ for fully ionized models. This means that BDM models with $\\Omega_0\\approx 0.15$ and reionization at  $z\\approx 1000$ are strongly constrained even in this very conservative case, and can survive the $y$ test only if most of the baryons form BDM around the reionization epoch. ",
+        "introduction": "Recombination of the primeval plasma is commonly assumed but was by no means inevitable. Theories exist that predict early reionization are as diverse as those invoking primordial seed fluctuations that underwent early collapse and generated sources of ionizing radiation, and models involving decaying or annihilating particles. The former class includes cosmic strings and textures, as well as primordial isocurvature baryon fluctuations. The latter category includes baryon symmetric cosmologies as well as decaying exotic particles or neutrinos.  The Compton $y$-distortion of the cosmic microwave background (CBR) provides a unique constraint on the epoch of reionization. In view of the extremely sensitive recent FIRAS limit of $y < 2.5\\times 10^{-5}$, we have reinvestigated constraints on the early ionization history of the intergalactic medium (IGM), and have chosen to focus on what we regard as the most important of the non-standard recombination history models, namely the primordial isocurvature baryon scenario  involving a universe dominated by baryonic dark matter (BDM), as advocated by Peebles (1987); Gnedin \\& Ostriker (1992) (hereafter ``GO\");  Cen, Ostriker \\& Peebles (1993) and others.  This class of models takes the simplest matter content for the universe, namely baryons, to constitute dark matter in an amount that is directly observed and is even within the bounds of primordial nucleosynthesis, if interpreted liberally, and can reconstruct essentially all of the observed phenomena that constrain large-scale structure. The BDM model is a non-starter unless the IGM underwent very early reionization, in order to avoid producing excessive CBR fluctuations on degree scales. Fortunately, early nonlinearity is inevitable with BDM initial conditions, $\\delta\\rho/\\rho\\propto M^{-5/12}$, corresponding to a power-spectrum $\\expec{\\delta_k^2}\\propto k^{-1/2}$ for the observationally preferred choice of spectral index (Cen, Ostriker \\& Peebles 1993).  Is it possible that the IGM has been highly ionized since close to the standard recombination epoch at $z\\approx 1100$? Perhaps the most carefully studied BDM scenario in which this happens is that by GO. In their scenario, $\\Oz=\\Omega_{b0}\\approx 0.15$. Shortly after recombination, a large fraction of the mass condenses into faint stars or massive black holes, releasing energy that reionizes the universe and heats it to $T>10,000\\K$ by $z=800$, so Compton scattering off of hot electrons causes strong spectral distortions in the cosmic microwave background. The models in GO give a Compton $y$-parameter between $0.96\\times 10^{-4}$ and $3.1\\times 10^{-4}$, and are thus all ruled out by the most recent observational constraint from the COBE FIRAS experiment, $y<2.5\\times 10^{-5}$ (Mather {\\etal} 1994).  There are essentially four mechanisms that can heat the IGM sufficiently to produce Compton $y$-distortions:  \\begin{itemize}  \\item Photoionization heating from UV photons (Shapiro \\& Giroux 1987; Donahue \\& Shull 1991)  \\item Compton heating from UV photons  \\item Mechanical heating from supernova-driven winds (Schwartz {\\etal} 1975; Ikeuchi 1981; Ostriker \\& Cowie 1981)  \\item Cosmic ray heating (Ginzburg \\& Ozernoi 1965)  \\end{itemize}   \\noindent The second effect tends to drive the IGM temperature towards two-thirds of the temperature of the ionizing radiation, whereas the first effect tends to drive the temperature towards a lower value $T^*$ that will be defined below. The third and fourth effect can produce much higher temperatures, often in the millions of degrees. The higher the temperature, the greater the $y$-distortion.  In the GO models, the second effect dominates, which is why they fail so badly. In this paper, we wish to place limits that are virtually impossible to evade. Thus we will use the most cautions assumptions possible, and assume that the latter three heating mechanisms are negligible. ",
+        "conclusions": "A reanalysis of the Compton $y$-distortion arising from early reionization shows that despite the radical sharpening of the  FIRAS limit on $y$, one still cannot conclusively assert that the universe underwent a  neutral phase. Non-recombining scenarios where the ionization is thermal, caused by IGM temperatures exceeding $10^4\\K$, are clearly ruled out. Rather, the loophole is for the dominant ionization mechanism to be photoionization. We have shown that for spectra characteristic of both QSO radiation and massive metal-poor stars, the resulting IGM temperatures are so low that typical CDM models with no recombination can still survive the FIRAS test by a factor of six. This conclusion is valid if the flux of ionizing radiation is not so extreme that Compton heating becomes important. This is not difficult to arrange, as the cross section for Thomson scattering is some six orders of magnitude smaller than that for photoionization.  For BDM models, the constraints are sharper. Non-recombining ``classical\" BDM models with $\\Omega_{igm}=\\Omega_0\\approx 0.15$ are ruled out even with the extremely cautious reheating assumptions used in this paper, the earliest ionization redshift allowed being  $z \\approx 130$. Such models involving early non-linear seeds that on energetic grounds can very plausibly provide a photoionization source capable of reionizing the universe soon after the period of first recombination inevitably generate Compton distortions of order $10^{-4}$. These include texture as well as BDM models, both of which postulate, and indeed require, early reionization ($z > 100$) to avoid the generation of excessive anisotropy in the cosmic microwave background on degree angular scales.  Thus BDM models with reionization at $z\\approx 1000$ can survive the $y$ test only if most of the baryons form BDM when reionization occurs, and are thereby removed as a source of $y$-distortion, at least in the diffuse phase. This may be difficult to arrange at $z>100,$ since once the matter is reionized at this high a redshift, Compton drag is extremely effective in inhibiting any further gas collapse until $z<100.$ Since it takes only a small fraction of the baryons in the universe to provide a source of photons sufficient to maintain a fully ionized IGM even at $z\\sim 1000,$ we suspect that most of the baryons remain diffuse until Compton drag eventually becomes ineffective. Moreover, the possibility that the IGM is only partially reionized at $z\\sim 1000$ ({\\it e.g.} GO), a situation which allows a lower value of the $y$-parameter, seems to us to be implausible as a delicate adjustment of ionization and recombination time-scales over a considerable range in $z$ would be required.  A complementary argument that greatly restricts the parameter space allowable for fully ionized BDM models appeals to temperature fluctuations induced on the secondary last scattering surface, both by first order Doppler terms on degree scales and by second order terms on subarcminute scales (Hu {\\etal} 1994). Thus, BDM models would seem to be in some difficulty because of the low limit on a possible $y$-distortion.   \\bigskip The authors would like to thank W. Hu, A. Reisenegger and D. Scott for many useful comments, and W. Vacca for providing stellar spectra. This research has been supported in part by a grant from the NSF.    \\clearpage"
+    },
+    "9307/gr-qc9307009_arXiv.txt": {
+        "abstract": "We study the power-law tails in the evolution of massless fields around a fixed background geometry corresponding to a black hole. We give analytical arguments for their existence at $scri_{+}$, at the future horizon and at future timelike infinity.  We confirm their existence with numerical integrations of the curved spacetime wave equation on the background of a Schwarzschild and a Reissner-Nordstr\\\"om black hole.  These results are relevant to studies of mass inflation and the instability of Cauchy horizons. The analytic arguments also suggest the behavior of the full nonlinear dynamics, which we study numerically in a companion paper. ",
+        "introduction": "In the study of nonspherical gravitational collapse, the late stages of black hole formation and nonspherical stellar dynamics, certain simplifications have been traditional. Typically, linearized perturbation theory has been used on a fixed background, and the results have been taken to be representative of nonperturbative collapse.  One of the most basic results \\cite{DeBr,RoWi,Ro} about the evolution of fields on a curved background is that ``sandwich waves'' are not usually possible.  At late times waves do not cut off sharply but die off in ``tails.''  In the context of perturbations of spherical objects, like stars or black holes, arguments have been given \\cite{Pr} leading to the conclusion that the tails have a specific power-law form. The intention of this paper is to analyze in detail in which regions of the spacetime this form for the tails holds and under what conditions they develop.  In section II we start by giving an analytical outline  of the development of tails in spherically symmetric fixed backgrounds. We present a somewhat more general and more pedagogical derivation of the results of the appendix of reference \\cite{Pr}. Moreover, we give two new results:  1) We show that for perturbations around a black hole, power-law tails develop not only at timelike infinity as was proven in \\cite{Pr} but also at $scri_{+}$ and at the black-hole horizon. This result is of relevance since the development of tails in these regions is crucial for the physics of mass inflation \\cite{HeKo} and the stability of Cauchy horizons. At least twice \\cite{CuMoPr,Le} in the literature it has been stated that the power-law tails are ``nonradiative,'' suggesting that the tails made no appearance at $scri_{+}$ or at the horizon.  2) Generalizing the arguments for the Schwarzschild background, we show that power-law tails develop even when no horizon is present in the background. This would mean, among other things, that power-law tails should be present in perturbations of stars, or after the implosion and subsequent explosion of a massless field which does not result in black hole formation.  In sections III and IV we confirm the first of these results for the Schwarzschild background by performing numerical integrations of the perturbation equations for different initial shapes and different multipole moments of the field. We also confirm that the results extend to the case of Reissner-Nordstr\\\"om black holes; this is the first clear evidence that power-law tails actually develop for the background of direct relevance to mass inflation and stability of Cauchy horizons. Section III contains a brief discussion of our numerical method; section IV contains our numerical results.  In section V we make some final remarks, especially about possible implications, both for the behavior of test fields evolving on on a time-dependent stellar collapse or explosion background, and for the behavior of a (spherically symmetric) self-gravitating massless field in a collapse or explosion situation. We develop this subject in a subsequent paper.  Tails have also been found important in the detailed calculation of gravitational waveforms from the inspiral collapse of binary systems. The tail back reaction appear as a correction to the 3/2 Post Newtonian equations of motions \\cite{AKOP}. ",
+        "conclusions": "When the dust of an approximately spherical collapse has settled, and spacetime inside the future lightcone of the collapse has approached Schwarzschild, Minkowski, or a stellar interior spacetime, any massless fields that were present in the collapse still show ``tails\" that linger, decaying only as a power of time. In particular the $l$-th multipole moment of a massless test field decays like $t^{-(2l+P+1)}$ at fixed radius at large times, with $P=1$ if there is an $l$-pole moment present before the collapse and $P=2$ otherwise.  But power-law tails are also present on $scri_{+}$, where they decay like $u^{-(l+P)}$, and if a black hole has formed, on the horizon, where they decay like $v^{-(2l+P+1)}$. The amplitude of the tails can be calculated as a function of the initial multipole moment, and in its absence, as an integral over the radiation going out to infinity during the collapse.  Neither this amplitude, nor the exponent of the power law, depend upon the spin of the massless field in question, nor, if the black hole is charged, on its charge. We have shown the origin of these features in an analysis which is based on \\cite{Pr}.  The spin- and charge-dependent parts of the effective radial potential of a massless test field on a Reissner-Nordstr\\\"om background are of an order that can be neglected in this analysis.  We have checked numerically that the tails are indeed present on null infinity and the horizon, and are independent of the spin of the field and the black hole charge.  The fact that power-law tails are present on the outer horizon of a Reissner-Nordstr\\\"om black hole after a generic collapse situation is of crucial importance to the mass inflation scenario. It has, to our knowledge, never been demonstrated explicitly.  Finally, one decisive step of our analysis was a regularity condition, either on the horizon when a black hole formed in the collapse, or else at the center of spherical symmetry.  This argument generalizes to any kind of boundary condition posed at small radius, and strongly suggests that perturbations of massless fields outside any spherical background object should also have power-law tails with the powers given above. In a subsequent paper we report results for a closely related {\\em nonlinear} problem: the implosion of a shell of scalar field.  \\thanks  We wish to thank Josh Goldberg and Fritz Rohrlich for pointing out early references.  This work was supported in part by grant NSF PHY92-07225 and by research funds of the University of Utah. J.P. acknowledges hospitality and support from the Institute for Theoretical Physics at UCSB and the National Science Foundation grant PHY89-04035."
+    },
+    "9307/astro-ph9307040_arXiv.txt": {
+        "abstract": "The Cosmic Microwave Background (CMB) fluctuations at very small angular scales (less than $10'$) induced by matter sources are computed in a simplified way. The result corrects a previous formula appearing in the literature. The small scale power spectrum from cosmic strings is then calculated by a new analytic method. The result compares extremely well with the spectrum computed by numerical techniques (when the old, incorrect, formula is used). The upper bound on the string parameters derived from OVRO data is re-examined, taking into account the non-Gaussian nature of stringy perturbations on small scales. Assuming a conventional ionization history,  the bound is $\\gamma G\\mu < 11\\times 10^{-6}$, where $\\gamma^2$ is the number of horizon lengths of string per horizon volume. Current simulations give $\\gamma^2 = 31\\pm 7$.   \\smallskip \\noindent {\\it Subject headings:\\/} cosmology: cosmic microwave background --- cosmic strings ",
+        "introduction": "\\setcounter{equation}{0} It is now widely accepted that cosmic microwave background (CMB) fluctuations test the physics of the very early universe.  With the recent flurry of experiments  on various angular scales (\\ct{MeyChePag91}; \\ct{Smo+92}; \\ct{Gai+92}; \\ct{Mei+93}; \\ct{Gan+93}; \\ct{Sch+93}; \\ct{Gun+93}), we are becoming more able to eliminate theories of the origin of these fluctuations.  There are two main contenders in the competition to explain the origin of these fluctuations: quantum fluctuations in the metric during inflation (see for example \\ct{BraFelMuk92}), and various kinds of semiclassical dynamics after phase transitions in field theories (\\ct{Kib76}; \\ct{Zel80}; \\ct{Vil80}; \\ct{Tur89}; \\ct{BarVil89}; \\ct{BenRhi90}).  Both are able to generate a more or less Harrison-Zel'dovich  spectrum of density fluctuations, and both produce CMB fluctuations which are consistent with the COBE observations (\\ct{Smo+92}), as far as the theoretical uncertainties allow (\\ct{Pee82}; \\ct{BonEfs87}; \\ct{BenSteBou92}; \\ct{BenRhi93}; \\ct{PenSpeTur93}).   The physics is quite different in each case, and each suffers in different ways from theorists' prejudices. However, it is to be hoped that future measurements of the spectrum of CMB fluctuations on a wide variety of angular scales will be able to distinguish experimentally between theories.  This paper is concerned with calculating the very small angular scale fluctuations from cosmic strings (\\ct{BouBenSte88}). Strings are perhaps the most venerable of the theories based on the dynamics of field theories during and after phase transitions.  Despite predating the inflationary scenario, the theory has suffered from analytic intractability and is consequently comparatively undeveloped.  Early work on string seeded galaxy perturbations (\\ct{BraTur86}; \\ct{TraBraTur86}) was based on the first numerical simulations (\\ct{AlbTur89}) whose detailed results have not all been confirmed by the two subsequent groups to work on the subject (\\ct{BenBou89}; \\ct{BenBou90}; \\ct{AllShe90a}; \\ct{AllShe90b}). What is lacking is a good analytic understanding of the evolution of the string network, although progress is being made in this regard (\\ct{KibCop90}; \\ct{CopKibAus92}; \\ct{Emb92a}; \\ct{Emb92b}). It is the intention of this work to try and create an analytic approach to the calculation of CMB fluctuations from strings.  Although the results presented apply only to fluctuations on very small angular scales (less than about 10$'$), these are precisely the scales on which strings make their most distinctive contribution.  The philosophy behind the current approach is quite simple.  It is that statistical measures of the string network itself can be used to calculate the statistics of the CMB fluctuations.  One can in fact guess the form of simple string  correlation functions on general grounds, backed up by some intuition gained from numerical simulations.   In this paper, Stebbins' (1988) formula for small angle CMB anisotropies is rederived in a more direct way, with an error in his and previous versions of this work corrected (Stebbins 1993). Then, the two-point string correlators are used to calculate the very high frequency end of the power spectrum of CMB fluctuations.  The result is compared to a numerical computation of the spectrum by  Bouchet, Bennett and Stebbins (1988), using the old anisotropy formula. In view of the approximations made, it is gratifying that the current approach reproduces both the shape and the amplitude of the spectrum very well.  Limits on the string linear mass density are then re-examined. The results can only be trusted on very small angular scales, for which the best experimental limits currently available are derived from recent VLA observations (\\ct{Fom+93}) and from  OVRO (\\ct{Rea+89}; \\ct{MyeReaLaw93}). The best experimental geometry for finding string is the recent RING experiment at OVRO (\\ct{MyeReaLaw93}),  which consists of 96 overlapping double difference fields in a circle of radius $\\sim 2^\\circ$ around the North Celestial Pole. However, sky coverage has been increased at the expense of sensitivity, so the best limits on the r.m.s.~fluctuations still come from the NCP observations (\\ct{Rea+89}). Other experiments at larger scales have been used in the past to constrain the string scenario, although in view of the current uncertainty surrounding the predictions of strings these should not be regarded as reliable.  The plan of the paper is as follows.  Firstly, I present the rederivation of Stebbins' (1988) formula, for small angle microwave anistropies. I then use this formula to derive an expression for the CMB power spectrum, subject to some plausible assumptions.  These assumptions are  justified by comparison to the numerical work of  Bouchet, Bennett and Stebbins (1988),  henceforth referred to as BBS. I then use the derived correlation function to obtain an upper limit on a certain combination of string parameters. To do this a revised Bayesian analysis is performed on OVRO data, which models the non-Gaussian statistics of strings. The bound, presented in (\\ref{eLim}), is expressed as one on $\\gamma G\\mu$, where $\\gamma^2$ is the number of horizon lengths of string per horizon volume, for $\\gamma$ is not well determined as yet.  Furthermore, this combination appears in all calculations of string-induced perturbations, and so bounding it is more useful than simply bounding the string tension. ",
+        "conclusions": "\\setcounter{equation}{0} In this paper a better analytic understanding of the CMB fluctuations produced by strings has been arrived at, at least on small angular scales (less than $10'$).  There are a number of ingredients in the success of the approach, which unfortunately make its extension to larger scales difficult.  Principally, the Minkowski space approximation results in a very simple formula for the projected temperature pattern, which depends only on the positions and velocities of the strings on the backward light cone of the observer.  This can only be justified for fluctuations on angular scales less than a degree.  An encouraging success is that the Gaussian approximation for strings, in which only the two point string correlators are used, reproduces accurately the numerical spectrum computed by BBS.  Using the (corrected) theoretical spectrum, a limit (\\ref{eLim}) can be derived on $\\ga G\\mu$ from observational data on small angular scales, where $\\ga$ is the number of horizon lengths of string per horizon volume. Because of the corrected formula, this translates to a more stringent limit than that given in Bennett, Bouchet \\& Stebbins (1989), even when statistical and theoretical uncertainties are properly accounted for. We also find that the correlation function on very small scales is approximately $C(r) \\simeq C(0)\\exp(-r/\\ze)$, with $\\xi \\simeq 10'$ (see Eq.~\\ref{eCorFun}).  This is rather different from the Standard Cold Dark Matter form $C(0)/(1+r^2/2\\al^2)$, with $\\al \\simeq 10'$ (\\ct{BonEfs84}). In experimental papers, a Gaussian autocorrelation function is often used to derive limits on temperature fluctuations, which is a reasonable approximation to the SCDM correlation function.  However, when limits on cosmic strings are required, an exponential correlation function is to be preferred.  I am extremely grateful to Albert Stebbins for many useful discussions, and particularly for communicating the revised formula for temperature anisotropies.  I am also grateful to him and his collaborators David Bennett and Fran\\c{c}ois Bouchet for making available their data. I thank also Paul Shellard for discussing his and Bruce Allen's numerical results, and Ron Horgan for a factor $\\log_{ e} 10$. This work was suported by the SERC."
+    },
+    "9307/astro-ph9307029_arXiv.txt": {
+        "abstract": " ",
+        "introduction": "For over a decade, the steep slope of the faint field galaxy counts and their very blue colors have been explained as the result of mild evolution in the spectral energy distributions (SED's) of galaxies. The excess of counts above non-evolutionary (NE) models has been claimed to be a factor of two  by $B \\sim 20$ (Maddox \\etal 1990, Loveday \\etal 1992) or by $B \\sim 22.5$ (Colless \\etal 1990), and  even factors of 5 to 15 by $B \\sim 25$ (Tyson 1988). These blue counts have been claimed, even with evolution, to be incompatible with a flat ($\\Omega = 1$) Friedmann universe (Koo 1990, Guiderdoni and Rocca-Volmerange 1990) or with the observed near-infrared K band counts (Cowie \\etal 1993). Yet recent redshift surveys show that galaxies fainter than 20th mag exhibit redshift distributions close to that predicted by NE models (Broadhurst \\etal 1988, Colless \\etal 1990, Lilly \\etal 1991).  A commonly held belief is that dramatic revisions to the conventional view are needed. Some have revised the cosmology by adopting a non-zero cosmological constant, $\\Lambda$ (Fukugita \\etal 1990).  Others have proposed non-conservation of galaxy numbers, due to mergers (White 1989, Cowie \\etal 1991, Broadhurst \\etal 1992), a disappearing population of dwarf galaxies  (Cowie \\etal 1991, Babul and Rees 1992), or dwarfs which have faded substantially in recent times (Broadhurst \\etal 1988).  An alternative  view is that the uncertainties of local and distant field galaxy data and galaxy models preclude a convincing case for any exotic theories at this time. Instead, a  NE model generated by {\\it trial and error} adjustments of galaxy luminosity functions (LF's) versus color is claimed to provide moderately good fits to the known data (Koo and Kron 1992). In this letter, we adopt a new {\\it objective} method to answer the question: what is the best fit that {\\it any} NE model can ever hope to make to the observations? ",
+        "conclusions": "Table 2 provides the resultant LF for each of the Bruzual color classes. Figure 1 compares various color-integrated LF's to some other recent derivations.  Our total LF to $M_{B_J} \\sim -18$ is seen to be a good match to the flat local LF derived by Loveday \\etal (1992). Though our predicted rise appears inconsistent with the faintest two points of the Loveday \\etal LF, Figure 1 shows that this steepening is compatible with recent LF's derived from fainter redshift surveys, including the {\\it local} LF derived by Eales (1993) and the faint field ($B_J > 20$) LF derived by Lonsdale and Chokshi (1993).  Our LF's divided by color can be directly compared to those adopted by Metcalfe \\etal (1991); our LF when summed over the same color intervals of $B-V$ (see Figure 1) also show a steep low-luminosity rise for the blue galaxies and consistent normalizations brighter than the valid limits of the Metcalfe \\etal (1991) fits.  Using these LF's, the predicted NE counts, colors, and redshift distributions are compared to existing observations and displayed in Figures 2 to 4, respectively. The fits are remarkably good and considerably improved over the hand-made NE model of Koo and Kron (1992). Though not shown, we also checked that the color-redshift distributions are also consistent, since in principle, different correlations of color-redshift may result in the same color or redshift marginal distribution.  Our new NE model shows only a slight deficit (less than a factor of two) of blue counts compared to the observations, even to the faintest reliabele limits of $B \\sim 25$.  The fit to the red ($r$) counts is excellent and the fit to the near-infrared ($K$) counts is acceptable.  The spectral energy distributions predict $V-K$ colors for elliptical galaxies that are too blue by approximately 0.3 mag compared to observations (see Fig. 11, Bruzual and Charlot 1993).  A shift of 0.3 mag towards brighter magnitudes in K for the fit to the K-band counts would actually improve the fit.  The color distributions are also very well matched from $B \\sim 19$ to $25$. It is intriguing to note that the comparison of the color distributions to the NE model shows that the {\\it excess faint galaxies are redder than average, not bluer}, contrary to standard lore based on previous NE models. The match to the redshift distributions is generally not as good fainter than about 21st mag, beyond which the predictions yield more low redshift $z < 0.1$ (i.e. local low-luminosity dwarfs with $M_{\\rm B} \\sim -12$ to $-17$) galaxies than observed, typically by about 10\\% of the total sample.  Whether this slight discrepancy can be explained by selection effects against compact or low surface-brightness dwarfs in the observations, by systematic errors in the zero-points of the faintest color data, or by the limitations of our model remains to be explored. For $B = 23$ to $24$, our new NE model predicts that about 33\\% of the galaxies will have redshifts $z \\ge 0.7$. This is greater than the 15\\% observed in the most recent data of Colless \\etal (1993), but not at a high level of statistical significance given the sparse redshift sample at these very faint magnitudes.  Despite the success of our near-optimal NE model, we are not arguing against some evolution. To the extent that the observed data are all accurate and our SED's are representative of real galaxies, our experiment indicates that no NE model exists that  will fit the observations totally, especially the higher (by $\\sim 40\\%$) counts by $B_J \\sim 24$. Whether these discrepancies can be explained by improved data and/or various evolutionary models will be explored in future papers along with more detailed error analysis and multicolor data.   Since our NE model is able to fit the counts, colors, and redshifts so well over a large range in magnitude, we explored why previous models predict fewer faint galaxies and redder colors.  Most assumed one-to-one conversions of galaxy morphology to color rather than including a dispersion of colors for each galaxy type; this assumption naturally results in too many red galaxies compared to reality. Others adopted a single LF shape for all galaxy types; with a standard Schechter LF, e.g., the blue galaxies tend to be underestimated at the extremes of the LF. Many models of the bluest galaxies included a characteristic luminosity for either a Schechter or Gaussian LF that was fainter than that for redder galaxies; this tends to underestimate the number of luminous blue galaxies, which are known to exist locally (specific examples can be found in the study of Markarian galaxies by  Huchra 1977). Some models even had the bluest galaxies as those with constant star-formation rate and thus $B-V \\sim 0.45 - 0.55$; many galaxies are bluer (again, see Huchra 1977).  In conclusion, we confirm our contrarian view from Koo and Kron (1992) that exotic theories are not yet required to explain existing faint field galaxy data.  Instead, we suggest that the local LF's for different color classes of galaxies adopted by previous studies are likely to be significantly in error.  We derive a plausible set of LF's by an objective technique.  The resulting NE predictions match well enough to the observations so that adoption of mild luminosity evolution remains a viable path to improved fits. Since the observational evidence for non-conservation of galaxy numbers with recent lookback time is now far less compelling, distant field galaxies may be resurrected as promising probes of the curvature of space."
+    },
+    "9307/astro-ph9307005_arXiv.txt": {
+        "abstract": "We present new ground-based data following up on the HST discovery of low-redshift Lyman $\\alpha$ absorption in the sight-line to the quasar 3C273. Our goal is to investigate the relationship between the low-column-density absorbers and higher column-density objects such as galaxies or H~II regions. Narrow-band filter observations with a coronograph show that there are no H~II regions or other strong H$\\alpha$ line-emitting gas within a 12 kpc radius of the line-of-sight to the quasar, at the velocities of three of the absorbers. Broad-band imaging in Gunn r shows that there are no dwarf galaxies at Virgo distances with absolute magnitude above M$_{\\rm B}\\approx$-13.5 and within a radius of 40 kpc from the line-of-sight to the quasar.  Finally, we present fiber spectroscopy of a complete sample of galaxies within a radius of 1{\\deg}, down to an apparent magnitude of B$\\approx$19. Analysis of this sample, combined with galaxies within 10 Mpc of the quasar line-of-sight taken from the literature, shows that the absorbers are definitely not distributed at random with respect to the galaxies, but also that the absorber-galaxy correlation function is not as strong as the galaxy-galaxy correlation function on large scales. We show that our data are consistent with the hypothesis that all galaxies more luminous than 1/10 L$^*$ have effective cross-sections (for association with absorbers whose neutral-hydrogen column-density (Log(NH)) is $>$13.0), of between 0.5 and 1 Mpc.  We also show a clear case of a Lyman $\\alpha$ absorber which has no galaxy brighter than M$_{\\rm B}$=-18 within a projected distance of 4.8 Mpc, and discuss the possibility that Lyman $\\alpha$ absorbers are destroyed in a rich galaxy environment. ",
+        "introduction": "\\label{introduction}  Understanding the origin and evolution of structure in the universe remains one of the most fundamental and active challenges of current astrophysical research. As the evidence in favor of a cosmological origin for the narrow, displaced absorption lines in QSO spectra became overwhelming, it also became clear that both the metal-line systems and the Lyman $\\alpha$ systems are invaluable tools for the study of some aspects of this problem. Since ground-based Lyman $\\alpha$ studies refer only to redshifts $\\ga$1.6, they complement studies of galaxy clustering properties, the majority of which involve redshifts much less than this.  However, precisely because the redshift regimes have been so different and because it has not been at all clear what relation exists between the typical low-column-density Lyman $\\alpha$ absorbers and galaxies, these two approaches have remained disjoint. It was somewhat unexpected, but pleasing, that low-redshift Lyman $\\alpha$ absorbers were found in sufficient numbers to enable meaningful  studies of the evolution of the Lyman $\\alpha$ absorbers and their relation to galaxies (\\cite{mor91,bah91b}). This has presented astronomers with the opportunity to join these two lines of investigation.  There are two levels at which such such attempts can be carried out: 1) Purely statistical investigations aimed at comparing the clustering properties of galaxies and Lyman $\\alpha$ absorbers, and 2) Investigation of individual cases in which the possibility of establishing the presence (or absence) of a clear link between a Lyman $\\alpha$ absorption line and something we could call a ``galaxy'' presents itself.  Preliminary discussions along these lines may be found in papers by \\cite{bah92a,bah92b} and by \\cite{sal92}. The present paper is a first attempt to pursue both these approaches along the sightline to 3C273. The remainder of this paper is organized as follows: In \\S~\\ref{observations} we describe the different sets of observations we have assembled to investigate the environment of the Lyman $\\alpha$ absorbers along the 3C273 sightline. In \\S~\\ref{analysis} we analyze them for possible associations or lack of associations of individual Lyman $\\alpha$ absorbers  with galaxies, and also give some statistical analysis of the clustering properties of the Lyman $\\alpha$ absorbers with galaxies. In \\S~\\ref{sumdisc} we discuss these results in light of current models for the Lyman $\\alpha$ absorbers and provide a brief summary and suggestions for further work.  Throughout this paper H$_0$ is taken to be 80 km/s/Mpc, the distance to the Virgo cluster is taken to be 16.0 Mpc (a distance modulus of (m-M)$_{\\rm Virgo}$=31.02) (\\cite{jac92}), and q$_0$ is taken to be 0. ",
+        "conclusions": "\\label{sumdisc}  We have assembled several types of observations in an attempt to find objects with which the Lyman $\\alpha$ absorbers along the line of sight to 3C273 might be associated, and in order to carry out statistical tests of galaxy-Lyman $\\alpha$ absorber association. In particular, we obtained narrow-band images centered on and off the expected position of any H$\\alpha$ emission which might be associated with three of the low redshift Lyman $\\alpha$ absorbers, and obtained deep broad-band images of a 17{\\arcmin}$\\times$17{\\arcmin} field centered on 3C273. Both these searches were negative.  Our failure to identify any broad-band or H$\\alpha$ emission from plausible ``galaxy-like'' objects a few tens of kpc from the 3C273 sight line at the approximate distance of the Virgo cluster will be checked by more sensitive and extensive searches for H$\\alpha$ by T.  Williams (private communication) and 21 cm emission by \\cite{vang93a} and \\cite{vang93b}.  We have also obtained redshifts for a large number of galaxies in the vicinity of the 3C273 sightline. Again, we find no unambiguous instance of association of any of the Lyman $\\alpha$ absorbers with individual galaxies. We define a number of samples for both the Lyman $\\alpha$ absorbers and the galaxies and estimate the 3-dimensional separation between each galaxy-galaxy pair and each Lyman $\\alpha$ absorber-galaxy pair based upon two models for converting the observed redshift difference between any pair into a radial separation, viz. i) the assumption of a pure Hubble flow and ii) a statistical model of ``perturbed'' Hubble flow based upon work of \\cite{dav83}. The resulting data base is used to carry out statistical tests to confirm or reject two null hypotheses about the association of galaxies and Lyman $\\alpha$ absorbers, namely: i) The Lyman $\\alpha$ absorbers show no tendency to cluster around galaxies ii) The Lyman $\\alpha$ absorbers cluster around galaxies exactly as the galaxies cluster about each other.  While neither of these two hypotheses can be unambiguously rejected in the sense that every combination of samples and flow hypotheses reject both of them at significant levels, the evidence from these tests, and from the galaxy-galaxy and Lyman $\\alpha$ absorber-galaxy two-point correlations themselves, points quite strongly to the conclusion that both hypotheses are false.  In particular, over length scales from about 1 to 10 Mpc there seems little doubt that the Lyman $\\alpha$ absorbers cluster around galaxies less strongly than the galaxies themselves cluster.  This is born out by an examination of a redshift interval centered at about z=0.078 at which a strong concentration of galaxies occurs but in the neighborhood of which there are no Lyman $\\alpha$ absorbers.  Additionally, we find at least one Lyman $\\alpha$ absorber for which no galaxy with absolute magnitude brighter than about -18 can be found closer than about 5 Mpc.  Taken together, all this evidence suggests that the most significant conclusion we have reached is that {\\it the majority of low-redshift Lyman $\\alpha$ absorbers are not intimately associated with normal luminous galaxies.}  In view of the fact that it has long been realized that at high redshifts there is an  absence of power in the  Lyman $\\alpha$ absorber two-point correlation function in redshift space, except possibly at the very smallest velocity separations, this conclusion is not too surprising. On the other hand, it is also fairly clear that there is {\\it some} tendency for the Lyman $\\alpha$ absorbers to cluster around galaxies, and even weak evidence that this clustering becomes strong at very small separations. Also \\cite{bah92a} have investigated the auto-correlation function of the low-redshift absorbers seen in the line-of-sight to H1821+643, and show that there is only a 4\\% probability that the observed `clumping' arose from a randomly distributed sample.  None of the foregoing points unambiguously, in our estimation, to a particular model for the formation and evolution of the Lyman $\\alpha$ absorbers. As have others, we simply offer the following speculations which appear to be compatible with the facts as they are presently understood:  At high redshifts, the Lyman $\\alpha$ absorbers consist primarily of entities which are only very loosely associated with larger mass objects (e.g. proto galaxies), and which are evolving fairly rapidly. Possibly this dominant population consists of absorbers in which gravitational binding  (eg by dark matter) plays no significant role. In addition to this group, there is a smaller population of absorbers which are evolving less rapidly, possibly stabilized by dark matter and which are clustered more strongly about galaxies. At very low redshifts, this latter population is beginning to constitute a large enough fraction of the absorbers that power in the two-point correlation function is detectable. Thus, the present mix of Lyman $\\alpha$ absorbers appears to have clustering properties intermediate between present-epoch normal galaxies and a random non-clustered population. This property also appears to be shared by the low-luminosity moderate redshift ``blue galaxies'' (\\cite{pri92}), leading to the plausible conjecture that the Lyman $\\alpha$ absorbers are more closely related to low mass, low luminosity galaxies than they are to L$^*$ galaxies, though the relation is clearly not one-to-one.  We have no good way at present of estimating the characteristic scale or masses of the low-redshift Lyman $\\alpha$ absorbers. A guess at a diameter of 30 kpc is as plausible as any. In particular, consider a pancake whose diameter is 30 kpc and whose thickness is 10 kpc. In this case, at the present epoch, a hydrogen column-density of 10$^{13}$-10$^{14}$ cm$^{-2}$ normal to the face of the absorber, coupled with estimates for the present-epoch energy-density of ionizating radiation leads to a total gas mass of order 10$^7$ M$_\\odot$, but a mass of only a few hundred solar masses of {\\it neutral} hydrogen. It is of interest that the mass function of neutral-hydrogen gas clouds appears to be truncated below about 10$^8$ M$_\\odot$ (\\cite{wei91}).  As Maloney (1992,1993) has shown, the gas in a flaring galaxy with decreasing column-density will undergo a rather sudden transition along its face from being mostly neutral to mostly ionized, with the consequence that few if any contours with neutral-hydrogen column-density of order 10$^{18}$ cm$^{-2}$ are known.  Similarly, it is conceivable that, given the appropriate run of length scale with mass, a sequence of masses would have the property of making a sudden transition in the mass function of neutral-hydrogen starting at about 10$^{8}$ M$_\\odot$, leading to a dearth of objects with total HI masses for several orders of magnitude below this.  If these speculations have any connection with reality then one might expect to see some similarity in the clustering properties of the Lyman $\\alpha$ absorbers and low mass galaxies.  We are currently attempting to obtain redshifts of galaxies of lower luminosity along the 3C273 sightline in order to investigate this possibility."
+    },
+    "9307/astro-ph9307017_arXiv.txt": {
+        "abstract": "Early photoionization of the intergalactic medium is discussed in a nearly model-independent way, in order to investigate whether early structures corresponding to rare Gaussian peaks in a CDM model can photoionize the intergalactic medium sufficiently early to appreciably smooth out the microwave background fluctuations. We conclude that this is indeed possible for a broad range of CDM normalizations and is almost inevitable for unbiased CDM, provided that the bulk of these early structures are quite small, no more massive than about $10^8 M_{\\odot}$. Typical parameter values predict that reionization occurs around $z=50$, thereby suppressing fluctuations on degree scales while leaving the larger angular scales probed by COBE reasonably unaffected. However, for non-standard CDM, incorporating mixed dark matter, vacuum density or a tilted primordial power spectrum, early reionization plays no significant role. ",
+        "introduction": "The first quantitative predictions of cosmic microwave background anisotropies in cold dark matter (CDM)-dominated cosmological models recognized that reionization by rare, early-forming objects could play a role in suppressing temperature fluctuations on small angular scales (Bond \\& Efstathiou 1984; Vittorio \\& Silk 1984). Now that the COBE DMR experiment has detected fluctuations on large angular scales (Smoot {\\etal} 1992) at a level (within a factor of two) comparable to that predicted  by CDM models, it is especially relevant to examine whether reionization can affect the degree scale searches that are currently underway.  Cold dark matter models are generally characterized by a late epoch of galaxy formation. However, the smallest and oldest objects first go nonlinear at relatively large redshift. In this paper we investigate, for a wide range of CDM normalizations, power spectra and efficiency parameters, whether reionization associated with energy injection by early forming dwarf galaxies can reionize the universe sufficiently early to smooth out primordial CBR temperature fluctuations.  Although we go into some detail in the appendix to make estimates of a certain efficiency parameter, our overall treatment is fairly model-independent, and can be used as a framework within which to compare various photoionization scenarios. Our basic picture is roughly the following: An ever larger fraction of the baryons in the universe falls into nonlinear structures and forms galaxies. A certain fraction of these baryons form stars or quasars which emit ultraviolet radiation, and some of this radiation escapes into the ambient intergalactic medium (IGM) and ends up photoionizing and heating it. Due to cooling losses and recombinations, the net number of ionizations per UV photon is generally less than unity.  Apart from photoionization, early galaxies can also ionize the IGM through supernova driven winds, an ionization mechanism that will not be treated in this paper. Although such winds can ionize the IGM by $z=5$, early enough to explain the absence of a Gunn-Peterson effect (Tegmark {\\etal} 1993), the relatively low velocities of such winds makes them unable to distribute the released energy throughout space at redshifts early enough (by $z\\approx 50$) to measurably affect the CBR.  Our approach will be to first write the ionization fraction of the IGM as a product of a number of factors, and then discuss the value of each of these factors in more detail. Let us write \\beq{FirstFactorEq} \\v = \\fs\\fupp\\fion, \\eeq where $$\\cases{ \\v&= fraction of IGM that is ionized,\\cr \\fs&= fraction of baryons in nonlinear structures,\\cr \\fupp&= UV photons emitted into IGM per proton in nonlinear structures,\\cr \\fion&= net ionizations per emitted UV photon. }$$ Let us first consider the case where the UV photons are produced by stars, and return to the quasar case later. Using the fact that a fraction $0.0073$ of the rest mass is released in stellar burning of hydrogen to helium, we obtain \\beq{fuppEq} \\fupp \\approx 0.0073\\left({m_pc^2\\over 13.6\\eV}\\right) \\fH\\fmet\\fuv\\fesc, \\eeq where $$\\cases{ \\fH&= mass fraction hydrogen in IGM,\\cr \\fmet&= mass fraction of hydrogen burnt,\\cr \\fuv&= fraction of energy released as UV photons,\\cr \\fesc&= fraction of UV photons that escape from galaxy. }$$ We will take the primordial mass fraction of helium to be $24\\%$, {\\ie} $\\fH=76\\%$. Now define the {\\it net efficiency} $$\\fnet = \\fmet\\>\\fuv\\>\\fesc\\>\\fion,$$ and \\eq{FirstFactorEq} becomes \\beq{SecondFactorEq} \\v\\aet{3.8}{5}\\>\\fnet\\>\\fs. \\eeq The key feature to note about this expression is that since $3.8\\times 10^5$ is such a large number, quite modest efficiencies $\\fnet$ still allow $\\v$ to become of order unity as soon as a very small fraction of the baryons are in galaxies. As will be seen in the next section, this means that reionization is possible even at redshifts far out in the Gaussian tail of the distribution of formation redshifts, at epochs long before those when the bulk of the baryons go nonlinear. This appears to have been first pointed out by Couchman and Rees (1986). ",
+        "conclusions": "A detailed discussion of how reionization affects the microwave background anisotropies would be beyond the scope of this paper, so we will merely review the main features. If the microwave background photons are rescattered at a redshift $z$, then the fluctuations we observe today will be suppressed on angular scales smaller than the angle subtended by the horizon at that redshift. This effect is seen in numerical integrations of the linearized Boltzmann equation ({\\eg} Bond \\& Efstathiou 1984; Vittorio \\& Silk 1984), and can be simply understood in purely geometrical terms. Suppose we detect a microwave photon arriving from some direction in space. Where was it just after recombination? In the absence of reionization, it would have been precisely where it appears to be coming from, say $3000$ Mpc away. If the IGM was reionized, however, the photon might have originated somewhere else, scattered off of a free electron and then started propagating towards us, so at recombination it might even have been right here. Thus to obtain the observed anisotropy, we have to convolve the anisotropies at last scattering with a window function that incorporates this smoothing effect. Typical widths for the window function appropriate to the last scattering surface range from a few arc-minutes with standard recombination to the value of a few degrees that we have derived here for early reionization models.  In addition to this suppression on sub-horizon scales, new fluctuations will be generated by the first order Doppler effect and by the Vishniac effect. The latter dominates on small angular scales and is not included in the linearized Boltzmann treatment because it is a second order effect. The current upper limit on CBR fluctuations on the 1 arcminute scale of $\\Delta T/T < 9\\tento{-6}$ (Subrahmanyan {\\etal} 1993) provides an interesting constraint on reionization histories through the Vishniac effect. In fact, according to the original calculations (Vishniac 1987), this would rule out most of the reionization histories in this paper. However, a more careful treatment (Hu {\\etal} 1994) predicts a Vishniac effect a factor of five smaller on this angular scale, so all reionization histories in this paper are still permitted.  The COBE DMR detection of $\\Delta T/T$ has provided a normalization for predicting CBR anisotropies on degree scales. Several experiments are underway to measure such anisotropies, and early results that report possible detections have recently become available from experiments at the South Pole (Meinhold \\& Lubin 1991; Shuster {\\etal} 1993) and at balloon altitudes (Devlin {\\etal} 1992; Meinhold {\\etal} 1993; Shuster {\\etal} 1993). There is some reason to believe that these detected signals are contaminated by galactic emission. Were this the case, the inferred CBR upper limits to fluctuations on degree scales might be lower than those predicted from COBE extrapolations that adopt the scale-invariant power spectrum that is consistent with the DMR result and is generally believed to be the most appropriate choice on large scales from theoretical considerations ({\\eg} Gorski {\\etal} 1993; Kashlinsky 1992). In the absence of such contaminations, the detected fluctuations in at least some degree-scale experiments are, however, consistent with the COBE extrapolation ({\\eg} Jubas \\& Dodelson 1993). The variation from field to field, repeated on degree scales, also may argue either for galactic contamination or else for unknown experimental systematics, or even non-Gaussian fluctuations. The results of other recent experiments such as ARGO (de Bernardis {\\etal} 1993), PYTHON (Dragovan {\\etal} 1993) and MSAM (Cheng {\\etal} 1993) have reinforced the impression that the experimental data is not entirely self-consistent, and that some form of systematic errors may be important.  The controversy over the interpretation of the degree-scale CBR fluctuations makes our reanalysis of the last scattering surface particularly timely. We have found that canonical dark matter, tailored to provide the 10 degree CBR fluctuations detected by the COBE DMR experiment, results in sufficiently early reionization (before $z \\approx 50$) over a fairly wide range of parameter space, to smooth out primordial degree-scale  fluctuations. Our middle-of-the-road model produces suppression by roughly a factor of two; it is difficult, although not impossible, to obtain a much larger suppression. This smoothing, because it is of order unity in scattering optical depth, is necessarily inhomogeneous. We predict the presence of regions with large fluctuations and many ``hot spots\" and ``cold spots\", corresponding to ``holes\" in the last-scattering surface, as well as regions with little small-scale power where the last scattering is more efficient. The detailed structure of the CBR sky in models with reionization will be left for future studies. Here we simply conclude by emphasizing that anomalously low values of $\\Delta T/T$ over degree scales are a natural corollary of reionization at high redshift.   \\appendix"
+    }
+}